Analysis of Swine Leukocyte Antigen Peptide …...Analysis of Swine Leukocyte Antigen Peptide...

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: May 29, 2020

Analysis of Swine Leukocyte Antigen Peptide Binding Profiles and the Identification ofT cell Epitopes by Tetramer Staining

Pedersen, Lasse Eggers

Publication date:2012

Document VersionPublisher's PDF, also known as Version of record

Link back to DTU Orbit

Citation (APA):Pedersen, L. E. (2012). Analysis of Swine Leukocyte Antigen Peptide Binding Profiles and the Identification of Tcell Epitopes by Tetramer Staining. Technical University of Denmark.

https://orbit.dtu.dk/en/publications/2c743356-d3d3-47d1-8aa9-27b199993014

Analysis of Swine Leukocyte Antigen Peptide Binding

Profiles and the Identification of T cell Epitopes by

Tetramer Staining

PhD Thesis Lasse Eggers Pedersen

2012



Tetramer Staining



1

„Think Big About Small Things”

Lasse Eggers Pedersen, September 2003

2

Supervisors:

Professor Gregers Jungersen

Section for Immunology and Vaccinology

National Veterinary Institute

Technical University of Denmark

Professor Søren Buus

Department of International Health, Immunology and Microbiology

Faculty of Health and medical Sciences

University of Copenhagen

Dr. William T. Golde

Foreign Animal Disease Unit

Plum Island Animal Disease Center, Agricultural Research Service

United States Department of Agriculture, United States Department of Homeland Security

Assessment committee:

Professor Peter M. H. Heegaard (chairperson)

Section for Immunology and Vaccinology

National Veterinary Institute

Technical University of Denmark

Professor Armin Saalmüller

Institute for Immunology

University of Veterinary Medicine Vienna

Professor Søren Skov

Laboratory of Immunology

Faculty of Life Sciences

University of Copenhagen

3

Front page artwork; Sketch animation of an SLA class I tetramer recognized by T cell receptors

at the cell surface, by Lasse Eggers Pedersen, December 2012.

Analysis of Swine Leukocyte Antigen Peptide Binding Profiles and the Identification of T cell

Epitopes by Tetramer Staining

PhD thesis 2012 © Lasse Eggers Pedersen

Table of Contents

Preface .............................................................................................................................................. 4

Acknowledgements .......................................................................................................................... 7

Summary ........................................................................................................................................... 9

Resumé (Danish summary) .............................................................................................................. 11

Abbreviations .................................................................................................................................... 13

List of Figures .................................................................................................................................... 15

1. Introduction .................................................................................................................................. 16

1.1. The Immune System ............................................................................................................... 16

1.1.1. Innate Immunity ............................................................................................................... 18

1.1.2. Adaptive Immunity ........................................................................................................... 19

1.2. MHC class I and The Swine Leukocyte Antigen ...................................................................... 20

1.3. Monitoring Cytotoxic CD8+ T Cell Frequency and Effector Function – Going from Mice and

Men to Cattle and Pig ............................................................................................................... 24

1.4. MHC class I Tetramers ............................................................................................................. 25

1.5. Foot-and-Mouth-Disease and Swine Influenza A Virus – Outbreak, Treatment and Control 26

1.5.1. Swine Influenza A Virus ................................................................................................... 27

1.5.2. Foot-and-Mouth-Disease Virus (FMDV) .......................................................................... 28

1.6. SLA Tissue Typing ..................................................................................................................... 33

2. Methodological Setup and Considerations ................................................................................. 35

2.1. Peptide-SLA Affinity Measures ................................................................................................ 35

2.2. Homogenous peptide Affinity and pSLA Stability Assays ........................................................ 36

4

2.3. Designing Negative Control Peptides for the Production and use of pSLA Tetramers .......... 39

2.4. Vaccination Trials against FMDV using a T Cell Specific Vaccine Construct ........................... 40

3. Paper I ........................................................................................................................................... 41

4. Paper II .......................................................................................................................................... 42

5. Paper III ......................................................................................................................................... 43

6. Paper IV ......................................................................................................................................... 44

7. SLA Tetramers ............................................................................................................................... 45

7.1. FMDV Tetramer Staining ........................................................................................................ 45

7.2. Influenza Tetramer Staining ................................................................................................... 48

8. Conclusions and Perspectives ...................................................................................................... 55

8.1. Simultaneous Analysis of Multiple Specificities ..................................................................... 57

8.2. Synthetic Immunogenic Protein (SIP) Vaccines ...................................................................... 58

8.3. Multimer Based Cancer Treatment ........................................................................................ 60

8.4. Answers lead to Questions ..................................................................................................... 61

Final Thoughts ................................................................................................................................... 62

Reference List.................................................................................................................................... 63

Preface

I clearly remember my first encounter with the Major Histocompatibility Complex (MHC) class I

protein. It was during the final year of my bachelor’s degree program in biochemistry and I was

attending a course in Protein Science. My group decided to work with the MHC molecule

because it was obvious that here was a protein with a striking functionality. What I did not

know was how it was related to immunology. In fact I did not know anything about immunology

at that time. Choosing the MHC molecule as our model in protein science turned out to be the

spark from which my passion for immunology burn strongly today.

After the course I was hooked on MHC, and decided to contact Professor Søren Buus, whom I

would later find was one of the world leading experts in this field. This was how I really got the

appetite for immunology and fortunately, Prof. Søren Buus agreed to let me into his Laboratory

of Experimental Immunology at the University of Copenhagen to do my bachelor’s, and later

5

my Master’s degrees in Biochemistry. For reasons that I do not know, Søren had just recently

put his interest into porcine MHC molecules, and hence he had his student (me) produce,

characterize and analyze Swine Leukocyte Antigen (SLA) proteins; The porcine MHC class I

molecules. Those three years of working with SLA introduced me to different areas of protein

science, biochemistry and not least immunology. Surrounded by extremely skilled and

professional people I was taught not only how to navigate in a laboratory but also how to

design, set up and adjust techniques to fit different experimental approaches on the way in

trying to solve scientific questions and problems that I have encountered, and hopefully will

continue to encounter throughout my career.

Graduating from the University of Copenhagen I found myself being a “still-wet-behind-the-

ears” but fairly well trained post Master’s scientist with a granted fellowship to go and work for

the United States Department of Homeland Security (DHS) and Department of Agriculture

(USDA). Through Prof. Søren Buus I got connected with Dr. William T. Golde and Prof. Gregers

Jungersen. These three scientists turned out to play a major role in my future research and

education as my PhD thesis mentors. A scientific collaboration forged between the Technical

University of Denmark (DTU), the University of Copenhagen (KU) and the USDA at Plum Island

Animal Disease Center (PIADC) in New York, USA, where I moved to and lived for a year and a

half to perform part of my thesis research. The rest of the research and work was carried out in

Copenhagen, Denmark at KU and DTU.

The research presented in this Ph.D. thesis was performed in the Section for Immunology and

Vaccinology, National Veterinary Institute, Technical University of Denmark, Copenhagen, at

the Institute of International Health, Immunology and Microbiology (ISIM), University of

Copenhagen, Denmark and at the United States Department of Homeland Security, Plum Island

Animal Disease Center, New York, USA in the period 2009 - 2012.

The project was funded by the Danish Council for Independent Research, Technology and

Production Sciences (274-09-0281) in collaboration with the Agricultural Research Service,

USDA (CRIS #1940-32000-053-00D) and the United States National Institutes of Health (NIH)

(HHSN266200400025C).

6

The thesis consists of eight chapters starting with an introduction to the field of immunology

seeing the MHC class I molecule as the focal point, followed by chapter 2 which describes

specific methodological approaches and other considerations. Chapters 3 – 6 are devoted to

the four manuscripts which constitute the backbone of this thesis. These chapters also include

respective results as well as general discussion. Chapter 7 discusses the SLA tetramer work

related to paper IV and influenza tetramer staining experiments performed at DTU. The eighth

and final chapter concerns general conclusions on the topic and thesis work as well as

perspectives.

This thesis includes four manuscripts (I – IV):

• Paper I Pedersen LE, Harndahl M, Rasmussen M, Lamberth K, Golde WT, Lund O, Nielsen

M, Buus S. Porcine major histocompatibility complex (MHC) class I molecules and

analysis of their peptide-binding specificities. Immunogenetics. 2011

Dec;63(12):821-34. doi: 10.1007/s00251-011-0555-3.

• Paper II Pedersen LE, Harndahl M, Nielsen M, Patch JR, Jungersen G, Buus S, Golde WT.

Identification of peptides from foot-and-mouth disease virus structural proteins

bound by class I swine leukocyte antigen (SLA) alleles, SLA-1*0401 and SLA-

2*0401. Animal Genetics. 2012 Sep 18. doi: 10.1111/j.1365-2052.2012.02400.x.

• Paper III Pedersen LE, Rasmussen M, Harndahl M, Nielsen M, Buus S, Jungersen G. An

analysis of affinity and stability in the identification of peptides bound by Swine

Leukocyte Antigens (SLA) combining matrix- and NetMHCpan based peptide

selection. Manuscript ready for submission to Immunogenetics.

• Paper IV Patch JR, Pedersen LE, Toka FN, Moraes M, Grubman MJ, Nielsen M, Jungersen G,

Buus S, Golde W.T. Induction of foot-and-mouth disease virus-specific cytotoxic T

cell killing by vaccination. Clinical and Vaccine Immunology: CVI. 2011

Feb;18(2):280-8.

Figure 1 illustrates the relation between the four manuscripts.

7

Figure 1. Diagram illustrating how the manuscripts of this thesis are related.

Acknowledgements

I have learned that great things are rarely achieved alone. For the past three years I have

crossed paths and thoughts with a lot of interesting people. My path has taken me from

Copenhagen to New York, dropping by Baltimore, further educating myself in Minneapolis,

learning about livestock viruses in Barcelona, presenting research in Tokyo, expanding my

horizon in Edinburgh before returning to Copenhagen. All the people that I have met on this

journey deserve a thank you for all the interesting discussions, for sharing of research ideas as

well as results, for opening their doors, for listening to opinions and for contributing to my

development and education as a scientist as well as a person.

I would like to thank my mentor and main advisor Prof. Gregers Jungersen for welcoming me

into his group at DTU, for making this project possible and for always being supportive and

taking his time to reflect on and help with ongoing projects and scientific problems.

My two co-advisors deserve special thanks as well;

Prof. Søren Buus for teaching me that serious work often produce serious results and for

reminding me that science is far from always easy, nor logic.

Paper I

Expression of SLA-1*0401 and the

establishment of peptide affinity

screening of porcine MHC class I comparing human and porcine β2m

Paper II

Determining the binding matrix for

SLA-2*0401 and the identification of

candidate T cell epitopes from FMDV for the SLA-1 and -2*0401 molecules

Paper III

Combining SLA matrix and NetMHCpan

predictions and high-throughput assays

to identify peptides bound by SLA-1, -2, and -3*0401 with high affinity and stability

Paper IV

Staining of FMDV specificCTLs and the

identification of a T cell epitope

using SLA tetramers and correlatingwith specifickilling

8

Dr. William Golde for letting me into his laboratory as well as into his family, for all the scientific

discussions and exciting experimental setups made possible by him and for putting an honor in

being my mentor.

A big thank you goes to my science buddy and good friend Michael Rasmussen. Thank you for

being critical upon my results and data analysis as well as for discussing interesting topics about

anything under the sun. Our Friday afternoons at Mikkeller’s brew house has not only been fun

but also inspiring.

I thank all the people in the Section for Immunology and Vaccinology, National Veterinary

Institute, DTU for their kind help in the laboratory as well as in regard to general issues and

problems, for advice and lunch time conversations, and for making me feel as part of a social,

diverse and very professional group of people, all connected by veterinary science. I owe Ulla

Riber, Dorte Fink Vadekær and Panchale Olsen a special thanks for their help with flow

cytometry, SLA typing and peptide affinity tests, respectively.

I also thank the Buus group at the Institute of International Health, Immunology and

Microbiology (ISIM), University of Copenhagen for always providing help and expertise when

needed. Particularly, I thank Mikkel Harndahl for his help with our high-throughput peptide

binding experiments as well as discussing the data output with me. The same goes for Morten

Nielsen of the Center for Biological Sequence Analysis, DTU, without whose contributions and

knowledge this thesis work would not have been what it is today.

All the people who are, or have been, part of the Golde group at PIADC in New York, during the

time that I have been working there deserve a warm thank you for always making me feel

welcome and as a part of the team, and for helping me with the continuous porcine tetramer

work that we have been performing on and off for the past three years. Special thanks goes to

Jared Patch for numerous discussions on different scientific topics, for escorting me in the BSL-3

facility before I got my Select Agent Clearance from FBI and for coming early and staying late on

the island with me whenever it was needed.

9

Last but not least, I would like to thank my friends and family. Without them none of this would

ever have been possible. I would like to thank my friends for sharing my enthusiasm, frustration

and laughs, and I would like to thank you, my mom and dad, for always believing in and

supporting me, for telling me how good I am and for never letting me doubt how proud you are

of me. A warm thank you goes to my sister Rikke as well, for always listening to what I have to

say and for sharing ideas. Finally, I would like to thank my wife, my love, my life, Kristina for her

limitless support and patience with me. You are always there when I need you the most.

It has been three intense years, which have required some sacrifice but also given me a

tremendous insight in the field of immunology as well as in myself. I am looking forward to

seeing all of you again now that my thesis is written and published.

Thank You

Copenhagen, December 2012.

Lasse Eggers Pedersen

Summary

The immune system has evolved in such a way that it is capable of specifically distinguishing

between self and non-self structures. Combining this with the ability to memorize and strongly

prepare for all pathogens that it has previously encountered, is essential for the proper and

effective function of the immune system, and provides the cornerstone for vaccine design. In all

vertebrate animals, CD8+ cytotoxic T lymphocytes (CTLs) survey the intracellular environment

for signs of invasion by pathogenic organisms such as viruses and bacteria. CTLs survey major

histocompatibility complex (MHC) class I molecules, which are highly polymorphic peptide

receptors which select and present endogenously derived peptides to circulating CTLs. Peptides

that are recognized by CTLs in the context of MHC are epitopes, and represent a small sample

of the pathogen proteome, making it possible for the immune system to specifically identify

and react upon non-self peptide fragments unique only to the foreign intruder. The

10

polymorphism of the MHC molecule effectively individualizes the immune response of each

member of any given species. Moreover, responding T cells recognize antigen ligands, only in

the context of peptide:MHC:β2m (pMHC) complex.

The gene encoding the MHC is one of the most polymorphic regions of the genome known.

Despite thousands of different human leukocyte antigen (HLA) variants identified, each

member of a species only inherits and expresses a few of these MHC alleles. The “MHC

fingerprint” of an individual can be identified by defining MHC alleles. This is classically called

“tissue typing” and is done by analyzing the reactivity of peripheral blood cells with sera unique

to MHC alleles. Such knowledge is paramount to analysis of the immune response regarding

MHC restriction and CTL recognition of pathogen-specific proteins.

Most of the polymorphism of MHC proteins is resident in the peptide binding groove. Hence,

each MHC is unique in the way it binds peptides, and by inference it individualizes the entire T

cell repertoire of the host. In this way, the diversity of MHC within a species makes immune

escape almost impossible for any intruding pathogen.

Characterization of the SLA class I and class II gene products and their peptide binding capacity

defines the T cell epitopes of any given pathogen proteome.

To date the analysis of MHC peptide interactions, strength (affinity) and stability of the

peptide:MHC complex, has been extensively reported in mice and humans, whereas data for

livestock animals such as the pig is rather limited. This thesis describes adapting and applying

analytical approaches, originally developed for man and mice to understand porcine MHC class

I peptide binding characteristics in relation to immune responses to vaccination or infection.

Applying proven technologies to newly produced, recombinant swine leukocyte antigen (SLA)

class I proteins yielded a body of data for peptide:SLA:β2m (pSLA) complex affinity and stability.

Mapping the SLA proteins for their peptide binding requirements along with the identification

of their cognate virally derived peptides, made it possible to explore the nature of the SLA

proteins and the roles they play in establishing adaptive immunity. The development of SLA

tetramers, enabled investigation of the specific CTL response elicited as a result of

immunization against foot-and-mouth-disease virus (FMDV) and swine influenza A virus. These

studies resulted in the identification of T cell epitopes from both viruses.

11

As SLA:peptide binding data accumulates in these and similar studies, it becomes possible for

collaborators at Center for Biological Sequence Analysis (CBS), DTU, to further strengthen the

NetMHCpan algorithm. This prediction tool now has the capacity for the selection of peptide

candidates to be bound by human (HLA), bovine (bovine leukocyte antigen (BoLA)) and swine

(SLA) MHC proteins.

Expanding the knowledge of porcine MHC class I molecules, mapping their peptide binding

matrices, and producing tetramers as well as improving today’s state-of-the-art method for SLA

peptide prediction will modernize immunological science of livestock species. These

approaches and results should lead to accelerated development of vaccines with increased

efficacy due to optimal activation of cell-mediated immune responses with minimal adverse

events.

Resumé (Danish summary)

Immunsystemet har udviklet sig på en sådan måde, at det er i stand til specifikt at skelne

mellem selv og ikke-selv strukturer. Ved at kombinere dette med evnen til effektivt at huske og

forberede sig imod samtlige patogener, som det tidligere er stødt på, er af afgørende betydning

for et korrekt og effektivt fungerende immunsystem og er selve grundstenen i ethvert design af

vacciner. I alle hvirveldyr overvåger kroppens CD8+ cytotoksiske T-lymfocytter (CTL'er) det

intracellulære miljø i søgen efter tegn på invasion af udefrakommende patogene organismer

såsom virus og bakterier. CTL'er overvåger major histocompatibilitets kompleks (MHC) klasse I

molekyler, som er stærkt polymorfe peptid-receptorer, der udvælger og præsenterer endogent

afledte peptider til cirkulerende CTL'er. Peptider, der genkendes af CTL'er i samspil med MHC

molekylet, er såkaldte epitoper og repræsenterer et lille fragment af det patogene proteom,

hvilket gør det muligt for immunsystemet specifikt at identificere og reagere imod fremmede

peptid fragmenter, som er unikke for den invaderende patogen. MHC molekylets polymorfi er

det, der individualiserer immunresponset for hvert enkelt individ i en given art. Endvidere

genkender kroppens T-celler kun antigene peptid ligander, når de fremstår som del af det

trimere peptid:MHC:β2m (pMHC) kompleks.

12

Genet der koder for MHC er kendt som en af de mest polymorfe regioner i genomet. Til trods

for eksistensen af tusindvis af forskellige humane leukocyt antigen (HLA) varianter, arver og

udtrykker hvert medlem af en art kun et par stykker af disse MHC alleler. Selve "MHC

fingeraftrykket" udtrykt af et individ, kan identificeres ved at definere MHC alleler. Den unikke

MHC allel kan fastslås ved en metode, traditionelt kendt som ”vævstypebestemmelse”, som

foretages ved at analysere perifere blod cellers reaktivitet overfor vævstypnings-sera. En sådan

viden er afgørende, for at kunne analysere et immunrespons i forhold til MHC restriktion og CTL

genkendelse af patogen-specifikke proteiner.

Størstedelen af MHC proteinets polymorfi er lokaliseret i den peptidbindende kløft. Derfor er

hvert enkelt MHC molekyle unikt i måden det binder peptid-antigener på og som følge heraf

individualiseres også hele individets T-celle repertoire via MHC. På denne måde er de enkelte

MHC molekylers forskellighed indenfor en bestemt art medvirkende til, at det stort set er

umuligt for indtrængende patogener, at undslippe immun systemets søgelys.

Karakterisering af individuelle SLA klasse I- og klasse II gen produkter og deres unikke peptid

bindende kriterier definerer T-celle epitoper fra ethvert givent patogen proteom.

Til dato er analysen af MHC klasse I molekylets evne til at binde og præsentere peptider, både i

form af bindingsstyrke (affinitet) og stabilitet af peptid:MHC komplekset, blevet omfattende

beskrevet i både mus og mennesker. Sådanne data er stadig relativt begrænsede for landbrugs

dyr såsom svin. Denne ph.d. afhandling beskriver hvorledes analytiske tilgange, der oprindeligt

er udviklet til mennesker og mus, kan tilpasses og anvendes i grise, for at opnå en bedre

forståelse af de porcine MHC klasse I molekylers peptid-bindings karakteristika i forbindelse

med immun responser rejst imod vaccinering eller infektion. En sådan anvendelse af tidligere

afprøvede analyse metoder sammen med nyligt producerede, rekombinante SLA klasse I

proteiner, førte til en betydelig mængde data for peptid:SLA:β2m (pSLA) kompleksets affinitet

og stabilitet. Kortlægning af SLA proteinernes bindingsmotiver samt identifikationen af

matchende virale peptider, gjorde det muligt at undersøge beskaffenheden af de forskellige SLA

proteiner samt de roller, de spiller i etableringen af adaptiv immunitet. Udviklingen af pSLA-

baserede tetramerer muliggjorde endvidere studier af specifikke CTL responser fremkaldt som

13

et resultat af immunisering mod mund- og klovesyge virus (MKS) og svine influenza A virus.

Disse undersøgelser resulterede i identifikationen af T-celle epitoper fra begge virus.

Med stadigt stigende mængder peptid bindings data for de forskellige SLA molekyler, opnået

igennem disse samt andres studier, er det muligt for vores samarbejdspartnere på Center for

Biologisk Sekvensanalyse (CBS), DTU, yderligere at styrke og forbedre den peptid forudsigende

algoritme NetMHCpan. Denne forudsigelses algoritme er nu i stand til at udpege peptid

kandidater til binding af både humane (HLA), kvæg (BoLA) og svine (SLA) MHC proteiner.

Ved at udvide kendskabet til porcine MHC klasse I molekyler, kortlægge deres peptid bindings

motiver og producere pSLA baserede tetramerer, samt forbedre nutidens aktuelle tekniske

niveau indenfor metoder til at forudsige SLA peptid binding, kan vi modernisere immunologisk

videnskabelige studier i landbrugs dyr. Disse fremgangsmåder og resultater må forventes at

føre til en fremskyndet og forbedret udvikling af vacciner med forøget effektivitet, på grund af

optimal aktivering af det cellemedierede immunrespons med tilhørende minimale bivirkninger.

Abbreviations

(In order of appearance)

MHC Major Histocompatibility Complex

SLA Swine Leukocyte Antigen

DHS Department of Homeland Security

USDA United States Department of Agriculture

DTU Technical University of Denmark

KU University of Copenhagen

PIADC Plum Island Animal Disease Center

ISIM Institute of International Health, Immunology and Microbiology

NIH National Institutes of Health

CTLs Cytotoxic T Lymphocytes

pMHC peptide:MHC:beta-2-microglobulin or peptide:MHC:β2m

HLA Human Leukocyte Antigen

14

FMDV Foot-and-Mouth-Disease Virus

CBS Center for Biological Sequence analysis

BoLA Bovine Leukocyte Antigen

DC Dendritic Cell

NK Natural Killer cell

Ab Antibody

Ig Immunoglobulin

Th T Helper cell

Treg Regulatory T cell

PRRs Pattern-Recognition Receptors

BCR B Cell Receptor

TCR T Cell Receptor

IL Interleukin

IFN Interferon

TAP Transporter associated with Antigen Processing

IPD Immune Polymorphism Database

ER Endoplasmic Reticulum

PLC Peptide Loading Complex

IFN Interferon

TNF Tumor Necrosis Factor

HA Hemagglutinin

NA Neuraminidase

ORF Open Reading Frame

VPg Viral Protein genome-linked

UTR Untranslated Region

IRES Internal Ribosome Entry Site

FMD Foot-and-mouth Disease

AA Amino Acid

UK United Kingdom

OIE Office International des Epizooties

15

DIVA Differentiating Infected from Vaccinated Animals

Ad5 Adenovirus 5

PRRSV Porcine Reproductive and Respiratory Syndrome Virus

ASFV African Swine Fever Virus

VSV Vesicular Stomatitis Virus

PCR-SSP PCR Sequence-Specific Primers

PBMCs Peripheral Blood Mononuclear Cells

ELISA Enzyme-Linked Immunosorbent Assay

LOCI Luminescent Oxygen Channeling Immunoassay

SPA Scintillation Proximity Assay

PK15 Porcine Kidney cell line 15

TCID50 Tissue Culture Infective Dose 50% pathological change

BoLA Bovine Leukocyte Antigen

APC Allophycocyanin

BV421 Brilliant Violet

EBV Epstein-Barr Virus

CMV Cytomegalovirus

SIP Synthetic Immunogenic Protein

Ad-SIP Adenovirus Vaccine Vectored SIP

CAF Cationic Adjuvant System

ISCOMs Immune Stimulatory Complexes

List of Figures

Figure 1: The relations between the four manuscripts of the thesis ................................................ 7

Figure 2: Key cellular components of the innate and adaptive immune system ............................. 17

Figure 3: Overview of cells and cell stages of the innate and adaptive immune responses ............ 18

Figure 4: The SLA class I molecule ..................................................................................................... 22

Figure 5: Illustration of an MHC class I monomer and tetramer ...................................................... 26

Figure 6: The influenza A replication cycle........................................................................................ 27

16

Figure 7: The foot-and-mouth disease virus virion and genome ...................................................... 29

Figure 8: Dark times .......................................................................................................................... 30

Figure 9: The foot-and-mouth disease virus replication cycle .......................................................... 32

Figure 10: SLA peptide affinity ELISA and peptide dose/response titration curve ........................... 36

Figure 11: The LOCI principle ............................................................................................................ 37

Figure 12: The SPA principle .............................................................................................................. 38

Figure 13: Tetramer staining of PBMCs from FMDV vaccinated swine ............................................ 46

Figure 14: Flow Cytometry plots of tetramer stained PBMCs from swine influenza A immunized

pigs .................................................................................................................................................... 50

Figure 15: SLA tetramer staining of porcine and human PBMCs ...................................................... 53

Figure 16: The SIP based vaccine principle ....................................................................................... 59

Figure 17: Indirect elimination of tumor cells by tetramers ............................................................. 61

1. Introduction

1.1. THE IMMUNE SYSTEM

In a world in which we are constantly exposed to pathogenic microbes and viruses it is

astounding how rarely we actually get ill. The primary reason for this is the immune system,

which consists of an array of different cells and mechanisms that form the body’s defense

system. The ability of the immune system to successfully overcome incessant pathogenic

invasions relies on its aptitude to sense and evaluate microbial threats leading to the control

and/or elimination of the intruding pathogen with minimal collateral host tissue destruction

and alteration of homeostasis (Goldszmid and Trinchieri 2012). Protection of the host against

the different responses of its own immune system is achieved by various levels of tight

regulation (Banchereau et al. 2012, Germain 2012, Murray and Smale 2012). The immune

system of vertebrates is often divided into two main arms; the innate and the adaptive immune

systems (Messaoudi et al. 2011). These systems rely on different immune cells generated in the

bone marrow from a common precursor (Figure 2), which display various functions. It is a

delicate balance between these two branches, and a coordinated action of their respective

17

cells, which underlies the immunologic function and success on which the host overall health

depends.

Figure 2. Key cellular components of the adaptive and innate branches of the immune system. Dendritic cells (DC),

natural killer cells (NK). The figure is adapted from (Messaoudi et al. 2011).

Traditionally, the immune response has been characterized by the type of antibodies (Ab)

produced and their ability to mediate a range of functions. Different antigens and differential

regulation of B cell stimulation, proliferation and differentiation into plasma cells (Shapiro-

Shelef and Calame 2005) leads to the production of Abs of various immunoglobulin (Ig)

isotypes. These different isotypes mediate effector functions such as neutralization,

complement fixation and opsonization of pathogens. With notable exceptions, the humoral

immune response is controlled by T lymphocytes that regulate B cell responses via the

cytokines they produce. This includes the T helper (Th) subsets, Th1, Th2, Th17, Th21, and the

CD4+/CD25+/Foxp3+ regulatory T cells (Treg). In addition to the effector function mediated by B

cells, a principal T cell effector function is mediated by the CD8+ cytotoxic T lymphocyte (CTL)

lineage. These cells are critical in anti-viral responses as they detect and kill virus infected cells.

A detailed overview of the different cells, cell stages and primary cytokines involved in the

induction and regulation of the innate and adaptive immune responses is illustrated in Figure 3

(Thakur et al. 2012).

18

Figure 3. Illustration of the different cells and cell stages involved in the induction and regulation of the innate and

adaptive immune responses. The figure was published in July 2012 by (Thakur et al. 2012), the author of this thesis

being co-author. For clarification, the primary focus of this thesis has been framed (dashed line square).

1.1.1. Innate Immunity

The innate immune response is the body’s first and primary line of defense against pathogenic

intruders, and provides immediate responses to a given infection, but without the

establishment of long lasting protection. The primary line of defense consists of mechanical

barriers like the gut and skin which are difficult for pathogens to penetrate.

The primary leukocytes mediating innate responses are the NK cells, mast cells, eosinophils and

basophils, also including the phagocytic cells; macrophages, dendritic cells and neutrophils

(Murphy et al. 2008). Two key functions of the innate immune system are the capacity to

broadly recognize and respond to conserved microbial products that are distinctive for non-

eukaryotic cells, mediated by germline encoded pattern-recognition receptors (PRRs), and

recruitment of adaptive immune cells to the site of infection by the induction and production of

cytokines (Janeway, Jr. 1989, Medzhitov 2007).

19

1.1.2. Adaptive Immunity

In contrast to the innate immune system, the adaptive immune system is not immediate, and it

takes time for specialized cells to be stimulated and proliferate in response to an infection.

Compared to the innate immunity that relies on germline-encoded receptors to detect the

presence of pathogens, the adaptive immune system utilizes a highly diverse set of antigen

receptors generated through somatic recombination of gene segments, thereby creating huge

numbers of B cell receptors (BCR) and antibodies with different specificities (Baranov and

Eppel' 1984, Tonegawa 1983). The antigen receptor on T cells, the T cell receptor (TCR), can be

nearly as diverse with specificities exceeding 1015 (Davis and Bjorkman 1988, Messaoudi et al.

2011, Schatz 2004, Tonegawa 1983). TCR diversity is dependent on inexact recombination of

gene segments during T cell development (Robins et al. 2009). The difference in antigen

recognition by BCRs and TCRs lies in the way B cells directly recognize antigens, whereas T cells

recognize antigens (non-self) in association with MHC (self), hereby combining the two

(Danilova 2012). Such MHC restriction is particularly important during the development and

maturation of primary lymphocytes in the thymus where T cells are forced to walk the fine line

of positive selection (Anderson and Takahama 2012, Bevan 1977, Kisielow et al. 1988). In the

bone marrow, auto-reactive B cells on the other hand, are negatively selected leading to

apoptosis (Gay et al. 1993, Hartley et al. 1993). B cell subsets are involved in the humoral

immunity by producing antibodies while T cells are responsible for the cellular immunity and

regulatory functions (Danilova 2012). TCR genomic rearrangements enable the adaptive

immune response to produce receptors capable of covering the extremely diverse pathogenic

pool by recognition of nearly any possible molecular structure.

During T cell development, positive selection ensures that T cells, expressing TCRs with self

MHC:peptide binding of a low level of affinity, receive sufficient signal to survive and further

develop. At this point T cells expressing a TCR stimulated by self MHC:peptide complexes, i.e.

auto-reactive, undergo negative selection resulting in induced apoptosis.

Mature, naïve T cells leave the thymus and start to patrol the lymphatic system of the host in a

constant search for pMHC complexes suitable for their specialized TCR. Upon encounter of a

20

pMHC complex presenting a foreign peptide antigen the T cell goes from being naïve to

becoming activated, undergoing proliferation and eventually exiting the lymphatic system in

high numbers. From there they migrate to the site of infection to clear the pathogenic invaders.

Once the immediate infection has been cleared most of the activated effector cells will die off,

whereas a minor subset of activated T cells will survive, and dependent on cytokine signals,

such as interleukin (IL)-12 or interferon (IFN)-α at distinct stages of the response, become

resting memory T cells (Mescher et al. 2006, Parish and Kaech 2009). These cells keep an

elevated precursor frequency of the activated T cells, capable of functioning as fast responders

to any subsequent similar infection.

1.2. MHC CLASS I AND THE SWINE LEUKOCYTE ANTIGEN

A large gene family exists in all vertebrates that encode the major histocompatibility complex

(MHC) cell surface molecules. These proteins are sub divided into three classes of molecules, I,

II and III. In humans, non-human primates and livestock animals, the focus is commonly set on

the MHC class I and II proteins. Also encoded in this gene locus are several other proteins such

as the transporter associated with antigen processing (TAP1 and TAP2), tapasin, MHC II DM and

DR, which are involved in the processing and load of antigen into the MHC binding groove

(Cresswell et al. 1999, Wijdeven et al. 2012).

The MHC class I proteins are encoded by a highly polymorphic gene cluster in most jawed

vertebrates (Ho et al. 2009c, Ujvari and Belov 2011). They are important protein determinants

in immune responses to infectious diseases and vaccines, as well as to autoimmunity and in

graft acceptance/rejection (Horton et al. 2004, Kelley et al. 2005, Singer et al. 1997). This region

was originally discovered through genetic studies of transplant incompatibility and the related

function of these proteins to create histocompatibility, or “self” identification through cell

surface expression of these highly polymorphic proteins (Gorer et al. 1948). The MHC class I

molecules are glycosylated, membrane-anchored proteins found on almost every nucleated cell

in vertebrates.

The MHC class I trimeric complex is made up of a heavy chain protein of approximately 43 kDa,

simultaneously non-covalently bound to the stabilizing subunit protein light chain, β2m of 12

kDa, and exogenous peptide (Cresswell et al. 1973, Grey et al. 1973). The heavy chain is made

21

up of three alpha domains, α1, α2 and α3 in addition to a membrane spanning domain and a

small cytosolic C-terminal tail (Figure 4). Most of the sequence variation is found in the α1 and

α2 domains which are responsible for the binding of exogenous peptides within the MHC class I

peptide binding groove, whereas the α3 domain and β2m are far more conserved expressing

homology to immune globulin domains (Bjorkman et al. 1987a, Bjorkman et al. 1987b, Orr et al.

1979). The higher conservancy of the α3 domain is most likely due to its role in the interaction

with the CTL surface expressed co-receptor CD8 (Salter et al. 1989).

In pigs the MHC is termed swine leukocyte antigen (SLA). The existence of the SLA complex was

clearly established more than 40 years ago (Vaiman et al. 1970, Viza et al. 1970), and it was

later determined that the SLA region extends for 2.0 Mbp on chromosome 7 (Chardon et al.

2000). Six different loci encode the classical (SLA-1, SLA-2, SLA-3) and non-classical (SLA-6, SLA-

7, SLA-8) SLA class I molecules (Ho et al. 2009b, Kusza et al. 2011). The SLA molecules are highly

polymorphic and extensive comparisons of the classical SLA class I and human MHC class I

genes (HLA-A, HLA-B, and HLA-C) have concluded, that there are more sequence homology

amongst the SLA genes than between any of these SLA loci and any HLA class I gene (Smith et

al. 2005). Furthermore, the overall genomic organization of the SLA class I region is quite

different from that of the HLA class I region (Lunney et al. 2009). As a result, the SLA class I

genes were assigned with numbers to avoid the implications that any loci were more

homologous to the genes of the HLA system. The non-classical class I genes are categorized so

because of their limited polymorphism and slight variations in the 3’ end specific of the

cytoplasmic tail compared to the three classical genes (Chardon et al. 2001). All SLA allele

sequences reported to date are publicly available at the Immune Polymorphism Database (IPD)-

MHC (http://www.ebi.ac.uk/ipd/mhc/sla/index.html).

22

Figure 4. Membrane bound SLA-1*0401 molecule illustrated with two different peptides bound (swine influenza A-

(blue) and Ebola virus-derived (green)). The α1, α2, and α3 domains are displayed in grey, orange and pink,

respectively, forming the heavy chain. The stabilizing subunit light chain, β2m, is shown in yellow. The membrane

spanning domain and cytoplasmic C-terminal (COOH) tail is illustrated by the dotted line. The figure is modified

from (Zhang et al. 2011).

The diversity in peptide ligand binding of the SLA proteins has been shown to influence swine

adaptive immune traits (Mallard et al. 1989a, Mallard et al. 1989b), vaccine responsiveness

(Lumsden et al. 1993, Rothschild et al. 1984) and disease resistance (Lacey et al. 1989, Madden

et al. 1990). Like the HLA molecules of humans, the SLA class I molecules of swine form a

binding groove which can be divided into six pockets responsible for binding of the individual

peptide antigen residues. These pockets are termed A, B, D, C, E and F and are responsible for

the binding of peptide residues P1, P2, P3, P6, P7 and P9, respectively (Madden 1995, Zhang et

al. 2011). The peptide specificity between different class I molecules is extremely diverse and is

defined by structural requirements such as the presence of certain amino acids in specific

positions of the peptide, especially at the peptide termini which account for most of the

peptide-MHC interactions by hydrogen bonds between the peptide backbone and the MHC

binding groove (Madden et al. 1992). These residues are often referred to as anchor positions

and define the binding motif of each individual MHC class I molecule.

23

For most MHCs, the critical residues for peptide binding are the P2 and P9 (or C-terminus if the

peptide is not a 9mer) of the peptide (Kubo et al. 1994). The side chains at P1, P4 and P5 of the

peptide generally protrude toward the solvent and away from the peptide binding groove

making them available for TCR recognition, whereas the P2, P3, P6 and P9 side chains face into

the groove and stay shielded from the TCR. The A and F pocket structures are well conserved

among many class I molecules. In comparison, those of the B through E pockets differ greatly

between heavy chains with respect to size and chemical nature (Matsumura et al. 1992),

sometimes leading to alternative peptide anchors in one of these pockets.

Degradation of cytosolic proteins predominantly by the proteasome, generates peptides of

both self and non-self origin depending on the status of the cell. Such peptides are actively

transported into the lumen of the Endoplasmic Reticulum (ER) by the TAP1 and TAP2 and

loaded onto the MHC class I in symphony with components of a multi molecular unit termed

the peptide loading complex (PLC) (Dalchau et al. 2011, Elliott and Williams 2005). Once a

peptide has been bound, the mature pMHC complex is being released from the PLC in the ER

and into the Golgi Apparatus from where it is transported to the cell surface. At the surface

pMHC complexes can be scanned by circulating CD8+ CTLs. In such a way, pSLA complexes

constantly signal the cell’s current state of integrity to the immune system of the host. Upon

recognition of foreign, non-self peptides, specific CTLs become activated to kill the infected or

transformed cell (Bevan 1995, Doherty and Zinkernagel 1975, Harty et al. 2000), even though a

given peptide:MHC complex is estimated to be present at the cell surface at 10 - 400 copies per

cell (Hunt et al. 1992).

The high number of SLA class I molecules and the level of polymorphism observed ensure the

diversity between different alleles. The number of MHC loci varies among species. Humans and

swine can express up to six different MHC class I molecules if they are heterozygous at all three

class I loci. However, the presence of duplicated loci leading to copy number variance of some

loci of classical SLA class I genes has been suggested to occur among SLA haplotypes (Ho et al.

2009b, Ho et al. 2010b). Such duplication leads to encoding of more than two class I alleles of a

gene in animals heterozygous for that particular locus (Ho et al. 2010b, Tanaka-Matsuda et al.

2009). Altogether, such SLA heterozygosity increases immune-competence in that a greater

24

variety of antigens are likely to be presented leading to a more robust immune response (Penn

et al. 2002).

1.3. MONITORING CYTOTOXIC CD8+ T CELL FREQUENCY AND EFFECTOR FUNCTION –

GOING FROM MICE AND MEN TO CATTLE AND PIG

Section 1.3 of this thesis is modified from section 3.2 in a paper previously published in Vaccine (Thakur et

al. 2012) with the thesis author as a coauthor including the responsibility for the writing of section 3.2.

In recent years the scientific fields of immunology and biochemistry have seen great advances

in the ability to detect and isolate antigen-specific CD8+ CTLs (Altman et al. 1996, Dunbar et al.

1998, Romero et al. 1998). Successive modifications of methods like the recombinant MHC

tetramer approach for staining of specific CTLs (Altman et al. 1996, Leisner et al. 2008), peptide

affinity ELISA (Sylvester-Hvid et al. 2002), cytokine sensitive ELISpot assays (Hutchings et al.

1989, Malyguine et al. 2007) and the use of chimeric molecules and transgenic animals (Choi et

al. 2002), have all contributed to the increasing amount of data describing cytotoxic immune

responses during infection and vaccination.

A major function of immune-activated CTLs is the exocytosis of cytotoxic proteins such as pore

forming proteins (perforins), serine proteases (granzymes), and cytokines such as IFN-γ, tumor

necrosis factor (TNF)-α and interleukin (IL)-2 (Zuber et al. 2005). Such components, revealing

the state of activation for individual CTLs, provide a relatively simple set of cytokines that can

be used to monitor and define both CTL frequency and function in responses against infections

that require T cells for protection (Seder et al. 2008). A potential drawback of using cytokines as

a measure for cell activation and function is that the secretion of many cytokines such as IFN-γ

is not necessarily indicative of CTL function, because they can be released by non-cytotoxic cells

involved in the innate and adaptive immune responses (Figure 3). An assay measuring the

secretion of molecules associated with lytic activity or the level of killing of specific target cells

by CTLs provides a better approach for assessment of cytotoxic T cell function. Present and

future approaches combining one such assay with highly specific MHC tetramer staining of CTLs

and additional flow cytometry-based intracellular staining assays are state-of-the-art and could

be applicable to models of almost any species of interest.

25

1.4. MHC CLASS I TETRAMERS

MHC tetramers have been described for mouse (Lemke et al. 2011), human (Altman et al. 1996,

Leisner et al. 2008) and bovine (Norimine et al. 2006). These complexes are four recombinant

MHC molecules loaded with synthetic peptides that are bound (usually through the Biotin–

Streptavidin interaction) to a fluorescent tag (Figure 5). They have been well known for more

than a decade (Altman et al. 1996), and being readily combined with functional assays they are

becoming more important as a tool in monitoring immune responses and in the enumeration

and characterization of T cells (Altman and Davis 2003, Wooldridge et al. 2009, Xu and Screaton

2002). Although some interaction of T cells with MHC monomers and dimers can be detected,

reproducible staining is best obtained with tetramers due to the higher avidity (Klenerman et

al. 2002). Up to three of the four MHC monomers in a tetramer can be recognized and bound

simultaneously by TCRs on the surface of the same antigen specific T cell thereby increasing the

tetramer-CTL interaction. Furthermore, tetramers themselves might bind each other to form

higher-order oligomers (Klenerman et al. 2002), which could further elevate the magnitude of

staining depending on the corresponding level of structural hindrance resulting from random

oligomerization. With the correct tetramers, antigen-specific CTLs can be directly enumerated

ex vivo from peripheral blood samples although in cases of very low in vivo frequencies in vitro

manipulations such as cell enrichment, cell expansion, and/or prolonged antigen stimulation

may enhance the individual frequencies for more detailed analyses (Addo et al. 2003, Kiepiela

et al. 2004, Leisner et al. 2008, Schmidt et al. 2009). However, such in vitro manipulations may

hinder the accuracy of monitoring the natural existence and frequency of CTL responses elicited

by infection or vaccination (Schmidt et al. 2009, thor Straten et al. 2000).

One main disadvantage of tetramers is their single-specificity which is a problem when

analyzing outbred animals and infections with certain viruses for which immunodominant

peptides do not exist. In cases where a complex set of epitopes might be targeted by T cells, it

would be necessary to perform analysis using several different class I tetramers produced with

several different peptide epitope candidates to obtain multiple specificities. Furthermore, in

the case of HLA and SLA, the limited number of recombinant MHC molecules available, and

scarce details of their peptide binding preferences, restricts the usage of MHC tetramers

(Saalmuller 2006).

26

Figure 5. The basic construction of an MHC class I monomer and tetramer recognized by TCRs at the surface of an

antigen-specific, CD8+ T cell. The fluorescently labeled streptavidin enables detection of CTLs that specifically bind

the pMHC tetrameric complex with higher avidity as compared to the lower affinity of the monomer. Figure

modified from (Klenerman et al. 2002).

Despite the established importance of virus-specific, CTL-mediated killing of virus infected cells

and the possibility to quantify induced CTLs by tetramer staining, the MHC heterogeneity of any

outbred population, and the exclusive specificity of the MHC tetramer with the corresponding

TCR, has so far made it difficult to use tetramer staining as an immunological correlate of

protection by validation of specific CD8+ CTLs (Thakur et al. 2012). Tetramer stained CTLs are,

however, important immune markers of the effects of vaccination as different vaccines or

vaccine strategies can be directly evaluated for their induction of peptide antigen-specific CD8+

CTL subsets. In addition, adoptive transfer of human CTLs specifically purified by tetramer

staining has been shown to induce protection against cytomegalovirus (Cobbold et al. 2005).

Furthermore, tetramers should allow the investigation of different T cell recognition patterns

with pMHC complexes for example by comparing peptides of different lengths.

1.5. FOOT-AND-MOUTH-DISEASE AND SWINE INFLUENZA A VIRUS – OUTBREAK,

TREATMENT AND CONTROL

The focus of this PhD project has been set on two of the most contagious and costly viral

infections that occasionally devastate the agricultural world, foot-and-mouth disease virus and

swine influenza A virus.

27

1.5.1. Swine Influenza A Virus

The swine influenza A virus strains are negative-sense, single-stranded RNA viruses of the

family Orthomyxoviridae (Beigel and Bray 2008). The most common influenza A virus strains

circulating worldwide are the H1N1, H3N2, and H1N2 distinguished by their divergence in the

two major glycoproteins hemagglutinin (HA) and neuraminidase (NA) (Brown et al. 1997,

Campanini et al. 2010, Olsen 2002, Richt et al. 2003). HA and NA display a fine balance playing

important roles in the influenza A replication cycle. HA mediates docking of the virions to their

target cells through binding to sialic acid (which is covalently linked to the terminal galactose of

an oligosaccharide, a glycoprotein or glycolipid). NA then functions by cleaving sialic acid from

galactose to avoid virion docking on non-respiratory epithelial cells as well as to release newly

formed virions from the surface of infected cells (Figure 6) (Beigel and Bray 2008, Jo et al. 2007,

Olsen 2002, Richt et al. 2003).

Figure 6. Replication of influenza A virus by initial docking of the virion to the surface of the target epithelial cell via

HA-Sialic acid-galactose interactions (1), followed by endocytosis, HA cleavage, and fusion of the viral and

endosomal membranes (2) which leads to the release and formation of ribonucleoprotein (RNP) particles into the

cell cytoplasm and further migration to the nucleus (3). Transcription and genome replication takes place within

the nucleus (4) resulting in translation of mRNA and the generation of nascent viral components such as HA and

NA (5), and culminating in the virion assembly and release at the cell membrane (6 + 7). The figure is modified

from (Beigel and Bray 2008).

Swine may play an important role in influenza ecology because they express viral receptors that

support infection with avian, swine and human influenza viruses, and by that act as mixing

vessel for new influenza A virus re-assortments, which may have the potential to be

28

transmitted to other hosts, including humans (Kida et al. 1994, Kundin 1970, Torremorell et al.

2012, Trebbien et al. 2011). Influenza A virus is mostly transmitted directly through pig-to-pig

contact and via aerosols. Seasonal influenza outbreaks greatly affect the meat and agricultural

industries every year due to illness, lowered meat production and in worst cases death within

animal herds.

The respiratory disease associated with influenza A virus has a global effect on swine as well as

birds and humans (Webster et al. 1992) leading to seasonal high demands of Oseltamivir - also

known as Tamiflu - which is the only currently available drug used in the prophylaxis of the

disease (Saxena et al. 2012). To avoid the risk of resistance and because of the greatly increased

price of the active compound of Tamiflu (shikimic acid), it is relevant to find alternative

approaches for the prevention and treatment of swine influenza, making it extremely

important to seek a better understanding of the immune response against influenza in swine,

as well as of the different mechanisms of transmission within pigs and across species

(Torremorell et al. 2012). Furthermore, in depth investigation and monitoring of porcine

immune responses elicited by new vaccines and by natural disease would contribute to

improved testing and development of future vaccines and vaccine delivery approaches.

1.5.2. Foot-and-Mouth Disease Virus (FMDV)

Foot-and-mouth disease virus is a member of the aphthovirus family of Picornaviridae. The

genome consists of a positive stranded RNA molecule which encodes a single polyprotein that

can be processed into four structural and nine non-structural proteins by virally encoded

proteases, including 3CPro and LPro (Guzman et al. 2008, Racaniello V.R. 2006, Stanway et al.

2005). The FMDV genome is surrounded by a densely packed icosahedral arrangement of 60

protomers, each consisting of the four structural proteins VP1, VP2, VP3 and VP4 encoded in

the P1 region (Figure 7) (Grubman and Baxt 2004, Rueckert and Wimmer 1984).

29

Figure 7. The FMDV virion (top) and the linear single-stranded RNA genome (bottom). The linear single-stranded

RNA genome is composed of a single open reading frame (ORF) encoding a polyprotein. The genomic RNA has a

linked viral protein (VPg) at its 5’ end. The long untranslated region (UTR) at the 5’ end contains an internal

ribosome entry site (IRES). The P1 region encodes the structural proteins (VP1 – VP4), whereas The P2 and P3

regions encode the nonstructural proteins associated with replication. The N-terminal leader protease (LPro

) is

encoded in addition to the 3C protease (3CPro

). This figure is modified from the Swiss Institute of Bioinformatics

(viralzone.expasy.org).

FMD is a highly contagious disease of cloven-hoofed animals including livestock such as cattle,

sheep, goats and swine. It is characterized by development of oral and pedal vesicles that result

in high morbidity but mortality is low. However, infection in young animals may be lethal

(Golde et al. 2005, Grubman and Baxt 2004, Pacheco et al. 2005). Devastating outbreaks such

as the one experienced in the United Kingdom (UK) in 2001 resulted in the slaughter of over 4

million animals and an estimated total loss to the UK economy of between US$ 12.3 and 13.8

billion (Scudamore and Harris 2002, Thompson et al. 2002).

Today, outbreaks that appear in previously FMD-free countries, such as in the UK, are

controlled by inhibition of animal movement followed by slaughter of infected as well as in-

contact animals (stamping out) and completed by decontamination (Pacheco et al. 2005). Such

methods are expensive in terms of lost agricultural production and export. In addition to this

there is also a considerable expense in the killing and burning of millions of animals, which

generally lead to public concern (Figure 8).

30

Along with the environmental problems raised by the need to eliminate these animals, such

factors underline the needs for alternative and better ways of controlling and dealing with

FMD. Although effective vaccines are currently available as chemically inactivated whole virus

antigen in adjuvant (Doel 2003), the use of vaccination is most often a decision of last resort.

The reason for this is that countries choosing to vaccinate animals (instead of stamping out) are

subjects to a prolonged loss of export markets for livestock and livestock products due to a

minimum of 6 months before regaining a status as FMD-free. Until recently this period was a

full 12 months in accordance with the previous version of the International Animal Health Code

of the Office International des Epizooties (OIE), which made the choice of vaccination further

unfavorable (Golde et al. 2005). In comparison, at present a country can get classified as FMD-

free by documentation of an absence of disease for 3 consecutive months when clearing an

outbreak by slaughter and decontamination.

Figure 8. Over four millions animals infected by FMDV were culled and burned during the 2001 outbreak in the UK

in an attempt to control and eradicate the virus. This picture was taken in North Cumbria, UK, one of the worst

affected areas in the country. Foto: Murdo Mcleod (www.guardian.co.uk).

Focusing more research on the control of FMDV and the development of further improved

vaccines and effective outbreak response programs, could ultimately lead to a state where

vaccination to prevent the disease, thereby saving millions of animals (and billions of US$),

would not just be a goal but an achievement.

Assays based on sensitive detection of antibodies against non-structural FMDV proteins, and

hence capable of differentiating infected from vaccinated animals (DIVA), have been developed

and are still being improved and redesigned for enhanced specificity and sensitivity (Brocchi et

31

al. 2006, Clavijo et al. 2004, Gao et al. 2012, Muller et al. 2010, Sorensen et al. 2005).

Furthermore, the OIE recommends that FMDV vaccines are composed of inactivated and

purified virus to remove non-structural viral proteins (Muller et al. 2010), thereby enabling

assay specificity for detection of non-structural protein-specific antibodies unique only to

infection and not to vaccination. One candidate for such a next-generation FMDV vaccine is the

“empty capsid” platform comprised of a replication-defective human adenovirus 5 (Ad5) viral

vector with the FMDV P1 capsid and 3CPro precursor inserted (Grubman et al. 2010).

Compatible with present diagnostic tests, this vaccine strategy has DIVA capability as well as

being effective in swine (Mayr et al. 1999, Mayr et al. 2001, Moraes et al. 2002) and cattle

(Pacheco et al. 2005). A limitation is seen for the vaccine (as for killed-virus vaccines and natural

infection in general) in that the neutralizing Abs raised against the vaccine strain of choice do

not provide cross-serotype protection. Due to such a requirement for vaccine efficacy, the lack

of cross-serotype protection limits the value of vaccination during outbreak or in disease

eradication (Rodriguez and Grubman 2009). Improved DIVA assays and vaccines would facilitate

the process of future vaccination programs to be preferred choices in the prevention, as well as

control of FMD outbreaks.

Previous work has demonstrated that vaccination of swine with wild type (3CPro intact) adeno-

vectored FMDV containing the strain A24 P1, 2A and strain A12 3C coding sequence (Ad5A24)

resulted in the generation of protective anti-FMDV antibodies (Moraes et al. 2002). These

experiments showed that swine vaccinated with a single dose of the Ad5A24 were completely

protected against a following challenge with FMDV. In comparison, the two control groups,

which were vaccinated with either a vesicular stomatitis virus G protein gene vectored by Ad5

(Ad5VSV-G) or with PBS alone, both developed characteristic signs of foot-and-mouth disease

(FMD). In addition, mutation of the 3CPro, specifically an amino acid change (C163S), resulted in

inhibition of the processing of P1 into mature capsid proteins and a reduced production of anti-

FMDV antibodies (Grubman et al. 1995, Patch et al. 2011). These data illustrate that 3CPro is

needed for successful structural capsid processing and virion assembly, which again is required

for proper induction of an anti-FMDV antibody response. A simplified illustration of the viral

FMDV replication cycle is shown in figure 9.

Figure 9. Illustration of the FMDV replication cycle. FMDV

RNA into the cytosol (2) leading to viral takeover of the cells own replication machinery and

viral proteins (3), as well as viral RNA

Endoplasmic Reticulum (4). New viral RNA forms the

newly forming capsids (5). Once assembled the new

membrane, causing aggregation, and eventually

particles capable of infecting healthy cells of the host

Development of vaccines focusing on the activation of

been somewhat neglected partly

infection, indicating a reduction of cell surface pMHC

al. 1998). Other reasons for CTLs

that neutralizing Abs have been demonstrated to be

disease (Mayr et al. 2001, Moraes et al. 2002)

appropriate assays for monitoring

more specific assays for the analysis of

effect of such CTLs in the clearance of virus from the infected host

CTLs to recognize peptide epitopes

antibodies makes them interesting candidates for enhancing FMDV immunity.

Illustration of the FMDV replication cycle. FMDV virus dock to the cell surface (

viral takeover of the cells own replication machinery and

as well as viral RNA within the replication complexes associated with new vesicles from the

iral RNA forms the genomes of virus particles which

. Once assembled the newly produced virus particles are

and eventually rupture of the cell (6), hereby releasing millions of new virus

capable of infecting healthy cells of the host (7). Figure illustration made by the author

vaccines focusing on the activation of an FMDV specific

partly due to observations done in the early events after FMDV

indicating a reduction of cell surface pMHC class I complex expression

reasons for CTLs being neglected in the context of FMDV infection

have been demonstrated to be central players in protection against the

(Mayr et al. 2001, Moraes et al. 2002), and technical difficulties in developing the

for monitoring such CTLs specific for FMDV. Despite this

more specific assays for the analysis of vaccine induced CTL responses could

in the clearance of virus from the infected host. Furthermore, t

CTLs to recognize peptide epitopes that would otherwise remain hidden to neutralizing


32

dock to the cell surface (1) and release the viral

viral takeover of the cells own replication machinery and the generation of new

ted with new vesicles from the

which become surrounded by

are transported to the cell

), hereby releasing millions of new virus

by the author.

FMDV specific CTL response have

observations done in the early events after FMDV

expression (Sanz-Parra et

neglected in the context of FMDV infection could also be

central players in protection against the

technical difficulties in developing the

Despite this, development of

could reveal a promising

Furthermore, the ability of

that would otherwise remain hidden to neutralizing


33

1.6. SLA TISSUE TYPING

Knowing which SLA alleles that are commonly occuring within a herd, a country or even on a

global scale can be of great importance in regard to vaccine development and the

establishment of immune protection in swine through broad coverage, highly specific, subunit

based vaccination against viruses such as the previously mentioned swine influenza virus and

FMDV. These technologies can also be applied to addressing other important viral diseases of

swine such as porcine reproductive and respiratory syndrome virus (PRRSV), African swine

fever virus (ASFV) and vesicular stomatitis virus (VSV).

Because of the extensive polymorphic nature of SLA genes, it is necessary to have very accurate

and sensitive typing methods to distinguish the different alleles (Ho et al. 2009c). Historically

the most important means for determining SLA class I antigen specificities was serological

typing using alloanti-sera (Kristensen 1988, Renard et al. 1988). This is compromised by the lack

of typing sera with documented specificities, limiting SLA characterization. Since MHC

molecules occasionally differ by sites inaccessible to antibody binding, and often share similar

(outer) epitopes, SLA typing sera are limited in identifying specific expression patterns. A

combination of sera is usually required to differentiate between animals, and most such typing

reagents are directed against an entire haplotype rather than individual alleles, further

complicating precise identification of particular SLA alleles. Hence the need for a more sensitive

and highly specific typing system, enabling the examination of SLA class I allelic diversities in

outbred populations, emerged.

Today, such needs are met by the application of DNA-based typing methods using low- and

high-resolution PCR sequence-specific primers (PCR-SSP). These allow high-throughput

identification of individual SLA class I alleles, and overall SLA haplotype screening of pig herds

and breeds of interest for class I (Essler et al. 2012, Ho et al. 2006, Ho et al. 2009c, Ho et al.

2009a, Martens et al. 2003) and class II alleles (Ho et al. 2010a). The principle of allele-specific

primer design is to identify primers where the 3’ end of one or both primers covers

polymorphic sites unique to their allele, thereby enabling the identification of that exact allele

using that particular pair of primers (Martens et al. 2003). To design site-specific primers, one

must know the DNA sequences of the different SLA alleles present in the swine herd or

34

population being studied. This limits the sequence-based typing approach to alleles with known

sequences. SLA allele sequences reported to date are publicly available at the IPD-MHC

(http://www.ebi.ac.uk/ipd/mhc/sla/index.html). As of November 2012, there are 131 class I

(SLA-1, SLA-2, SLA-3 (all classical class Ia), and SLA-6 (non-classical class Ib)) alleles officially

designated (Table 1) with a total of 31 class I high-resolution haplotypes assigned at the allele

level, which puts the SLA system among the most well characterized MHC systems in non-

primate species (Ho et al. 2012).

Table 1. Statistics of the current SLA sequence database displaying the number of newly identified alleles since

2005. The total number reported for each allele group is indicated in the lower graph. The table was modified and

used with approval from the corresponding author (Ho et al. 2012).

The relevance of SLA typing can be linked to the SLA peptide ligand binding and presentation

(Lunney et al. 2009). An initial step in identifying virally derived T cell epitopes, that generate a

protective T cell response against infection in pigs, is to identify which SLA alleles are expressed

by the animals and which peptides they bind and present. Once candidate peptides or proteins

have been identified, they may be used to specifically target, activate and/or monitor T cells

mediating immunity through subunit based vaccination or synthetically designed pSLA

2005

2009

Currently

Identified alleles

SLA class I 88

SLA class II 127

Previously

identified alleles

(2005)

Newly identified

alleles

Total

SLA class I 88 + 37 = 125

SLA class II 127 + 37 = 164

Previously

identified alleles

(2009)

Newly identified

alleles

Total

SLA class I 125 + 6 = 131

SLA class II 164 + 10 = 174

46 48

28

9

35

multimers. Highly specific vaccines based on detailed information about SLA peptide binding

and presentation could become useful in preventing the outbreaks of viruses such as swine

influenza and FMDV, hereby limiting the almost inevitable financial losses and animal suffering,

which are associated with such outbreaks.

2. Methodological Setup and Considerations

This chapter describes the methods and experimental approaches used throughout the PhD

project, but not explicitly described in the respective manuscripts. General and detailed

materials and methods sections for all methods and approaches of my PhD work can be found

in the individual manuscripts presented in this thesis, Paper I – IV.

2.1. PEPTIDE-SLA CLASS I AFFINITY MEASURES

In the studies performed throughout this PhD project the affinities of hundreds of different

peptides for binding by the three SLA class I molecules, SLA-1*0401, SLA-2*0401 and SLA-

3*0401 have been measured. The assay initially used for this was an enzyme-linked

immunosorbent assay (ELISA) previously developed for HLA (Sylvester-Hvid et al. 2002) and

through this project adapted to SLA (Paper I – III). The ELISA as such was first described in 1971

and is based on specific Abs capturing antigen which is then detected by an enzyme conjugated

detection Ab in symphony with a substrate to give a visible signal (Engvall and Perlmann 1971,

Van Weemen and Schuurs 1971). Today the use of monoclonal Abs with various specificities is

common practice in most laboratories allowing for an extremely broad range of antigen

detection measures. The ELISA used in this project was based on the capture of biotinylated

SLA molecules followed by the detection of human β2m using a monoclonal Ab (Figure 10A).

This assay principle is based on data showing the affinity between β2m and the MHC heavy

chain is enhanced in the presence of bound peptide ligand (Elliott et al. 1991). Hence β2m is

only stably folded with SLA class I when there is peptide binding. This results in no detection of

peptide-empty SLA molecules in the assay. Due to the lack of a monoclonal Ab specific for

porcine β2m we decided to build the assay with human β2m. This decision was based on the

findings that both human and porcine β2m show equal capacity to support the folding of pMHC

36

class I complexes irrelevant of the heavy chain expressing an HLA or SLA α3 domain as

presented in Paper I (Figure 4 + 5) of this thesis. The amount of peptide needed to fully saturate

half of the SLA molecules made available will give the affinity constant (KD) of each peptide

tested once correlated to a pre-folded HLA-A*02:01 standard curve based on non-linear curve

regression (Figure 10B).

Figure 10. Illustration of the SLA peptide affinity ELISA setup (A) and an example of curve-fitting to peptide

dose/response titrations, and the corresponding pSLA complex formation (B). The ELISA plate is covered with

streptavidin (orange) which bind the biotin (pink) linked to the SLA heavy chain molecule (green) with bound

peptide (red) and the co-bound stabilizing subunit protein β2m (purple). β2m-specific primary Ab (yellow) captures

correctly formed pSLA complex made visible by detection via an enzyme (glowing blue) conjugated detection Ab

(grey). The peptide dose/response curve illustrates a peptide (MTAHITVPY) bound with extremely high affinity (KD,

0.8 nM) as presented in Paper II of this thesis.

2.2. HOMOGENOUS PEPTIDE AFFINITY AND pSLA COMPLEX STABILITY ASSAYS

Luminescent oxygen channeling immunoassay (LOCI) is a technology which enables high

throughput, homogenous screening of peptide-MHC class I interactions, among a wide range of

other rapid, and quantitative determination possibilities. The technology itself was first

described in the nineties (Ullman et al. 1994, Ullman et al. 1996), and the assay consists of two

primary components, donor- and acceptor beads, the latter emitting a signal when brought into

close proximity with excited donor beads (Figure 11).

37

Figure 11. The LOCI principle. Biotinylated pSLA (green) is bound to a streptavidin coated donor bead (blue) and

captured by an anti SLA monoclonal Ab (PT85A, black) which is conjugated to an acceptor bead (orange).

The donor beads contain a blue-green colored photosensitizing agent, phthalocyanine, which

converts ambient oxygen into singlet oxygen upon illumination at 680 nm. The lifetime of

singlet oxygen in water is very short (4 µs), hence it can only travel approximately 200 nm

before returning to its ground state. This enables acceptor beads to become activated only

when brought into close proximity of the donor bead, resulting in a chemiluminescent signal via

the PT85A monoclonal Ab capture of correctly folded pSLA complex at the surface of the

streptavidin conjugated donor beads (Figure 11). The resulting light emission from activated

acceptor beads is at a higher energy than the light used for excitation, which result in very low

background signals and a high signal-to-background ratio. In addition to this, the beads are

bioactive and can easily be conjugated with Abs or other proteins (such as streptavidin).

Furthermore, the homogenous nature of the LOCI assay leaves out the washing steps needed

in-between steps for the ELISAs. Altogether, such assay qualities make the LOCI preferable to

use in present and future peptide-SLA affinity analyses. To measure the assay signal an

instrument like the EnVision® or EnSpire™ (Perkin Elmer Life Sciences) multilabel reader is

needed. The LOCI technology was adapted to screen for peptide:SLA affinities having

biotinylated heavy chain proteins and an SLA-specific Ab available and adjusted to the assay.

LOCI based peptide affinity analysis is described and presented for the SLA-2*0401 and SLA-

3*0401 molecules in paper III.

38

In addition to the affinity, which is a measure for the likeliness of a given peptide to be bound

by the MHC class I molecule, accompanying stability measures of pSLA complexes were

included. The stability of pMHC has been proposed to be a better determinant of

immunogenicity than the actual binding affinity (Busch and Pamer 1998, Harndahl et al. 2012,

Mullbacher et al. 1999, van der Burg et al. 1996). To measure such stabilities we used a

radiolabeled-based scintillation proximity assay (SPA). The first SPA emerged in the late 1970s

(Hart and Greenwald 1979a, Hart and Greenwald 1979b) making the technology commercially

available. By radio-labeling porcine β2m with the 125I isotope we were able to develop a high

throughput, homogenous assay for measuring pSLA complex stabilities. Such stability measures

are based on the energy transfer between β-particles from the iodine isotope and the

scintillation material imbedded in streptavidin coated FlashPlates which bind biotinylated

peptide:SLA:125I-β2m complexes. Upon peptide binding by the SLA molecule, 125I-labeled β2m

will co-bind, which will bring the β2m close to the scintillation material in the plate, hereby

activating it by the emitted electrons (β-particles). Activation of the plate then results in a

positive signal read-out (Figure 12).

Figure 12. Illustration of the SPA principle. All components are the same as described in Figure 10 with the

exception of the β2m now being labeled with 125

I (glowing pink) and the signal emitted from the activated

scintillation material (glowing blue).

The 125I isotope was chosen partly due to its short ranged electron emitters, hence only labeled

125I-β2m brought in close proximity of the FlashPlate scintillation material by the streptavidin-

biotin interaction, will lead to signal in the reader. Electrons emitted from unbound 125I-β2m will

be absorbed by the surroundings. The dissociation of 125I-β2m is used as a measure of the

stability of the pSLA complex. Excess of un-labeled β2m is added to the sample when steady-

state of the complex folding reaction is reached, and just prior to sample measurement. This is

39

done to ensure that dissociated peptide:SLA:125I-β2m complex is only re-associated by support

of un-labeled β2m, not leading to a false high stability by regeneration of measurable

peptide:SLA:125I-β2m complexes. Samples are measured continuously in a TopCount NXT liquid

scintillation counter for 24 hours at 37⁰C. The assay and the corresponding stability data

achieved for SLA-1*0401, SLA-2*0401 and SLA-3*0401 is presented and discussed in detail in

Paper III.

2.3. DESIGNING NEGATIVE CONTROL PEPTIDES FOR THE PRODUCTION AND USE OF

pSLA TETRAMERS

Critical to interpreting tetramer binding data is the capacity to measure background signal. In

order to address this, two peptides derived from the VP35 region of the Sudan Ebola virus and

from the Vibrio Cholera bacteria, respectively were used as nonspecific controls. The binding

prediction for the peptides with two SLA molecules, SLA-1*0401 and SLA-2*0401, and in vitro

ELISA affinity tests, revealed both peptides as being strongly bound as presented in Paper II.

The idea was to produce tetramers using these peptides as negative controls when staining

PBMCs. It was based on the presumption that animals under study had never encountered

either Ebola or Cholera, and therefore should not be expected to recognize such tetramers,

leading to positive staining of their CTLs. This approach was most successful using the Ebola

peptide, leading to relatively low false positive background staining. Using the Ebola peptide

construct as the reference tetramer was effective, as presented in details in Paper IV.

Considering the binding and overall conformation of pMHC complex, combining it with

recognition by the TCR being based on the peptide and MHC protein, it follows that the smaller

the AAs of the peptide, the less available would the peptide residues be for TCR identification.

Following this logic, better control peptides for tetramer formation could be designed resulting

in less false-positive tetramer staining. Synthetic 9mer peptides were designed based on three

criteria; (i) satisfying the residue preferences inherent in the SLA-1*0401 binding matrix anchor

positions P2 and P9, (ii) be comprised of as many alanines or glycines as possible, and (iii) be

ranked as strong binders by the NetMHCpan algorithm. Following these criteria, five such

peptides were synthesized and tested for affinity binding by SLA-1*0401. Three of the five

peptides had high binding affinities of 19, 40 and 104 nM. The two peptides with the highest

40

affinities were included in the tetramer experiments to provide a “null” tetramer. From those

studies we observed that the peptide ASYGAGAGY was consistent in exibiting low signal in the

flow cytometry analysis of T cell staining. This peptide is currently our best option as a negative

control in the tetramer experiments.

2.4. VACCINATION TRIALS AGAINST FMDV USING A T CELL SPECIFIC VACCINE

CONSTRUCT

Performing the vaccination and boosting trials presented in Paper IV we compared two

different vaccine constructs which differed in the functionality of the 3Cpro as described above.

The reason for using a 3Cpro mutated version (C163S) of the original vaccine construct was in an

attempt to specifically induce a CTL response. We hypothesized that a stronger CTL response

would be induced by such an “FMDV-T activating” vaccine, as opposed to the FMDV-B

activating construct displaying a functional 3Cpro protein. The hypothesis was based on

previously published data showing that the mutated and inactivated protease lacks the ability

to successfully cleave the FMDV P1 protein into the sub-proteins (VP1 – 4) which form the viral

capsid, hereby enabling the virus to leave the cell. Instead, viral protein would stay captured

inside the cells leading to proteasome degradation and elevated availability of various FMDV P1

derived peptides for MHC class I load in the ER, and presentation at the cell surface. The work

described in Paper IV presents the approaches made to test this hypothesis and the outcome of

the experiments.

41

3. Paper I

Porcine major histocompatibility complex (MHC) class I molecules and analysis of their

peptide-binding specificities

Pedersen LE, Harndahl M, Rasmussen M, Lamberth K, Golde WT, Lund O, Nielsen M,

Buus S.

Immunogenetics. 2011 Dec:63(12):821-34. doi: 10.1007/s00251-011-0555-3

The work presented in this manuscript was focused on the production of the SLA-1*0401

molecule and the porcine β2m stabilizing subunit protein. Generating chimeric human/swine

MHC-molecules by systemically transferring domains of the frequently expressed SLA-1*0401

onto a HLA-I molecule (HLA-A*11:01), enabled us to analyze the peptide-binding characteristics

of these molecules and validate the effects of pMHC complex formation supported by either

human or porcine β2m.

In this study we demonstrated that the approaches developed for the “Human MHC Project”

can be successfully applied to other species such as pigs. Furthermore, we here presented the

complete PSCPL mapped binding motif for one of the most common SLA molecules worldwide,

and showed that pMHC complex formation can be equally supported by human as well as

porcine β2m, regardless of the α3 domain of the heavy chain being of human or porcine origin.

In addition, we demonstrate the successful use of a pan-specific predictor of peptide:MHC class

I binding, the NetMHCpan, in predicting the specificities of the SLA-1*0401 molecule.

ORIGINAL PAPER

Porcine major histocompatibility complex (MHC) classI molecules and analysis of their peptide-binding specificities

Lasse Eggers Pedersen & Mikkel Harndahl & Michael Rasmussen & Kasper Lamberth &

William T. Golde & Ole Lund & Morten Nielsen & Soren Buus

Received: 25 March 2011 /Accepted: 20 June 2011 /Published online: 8 July 2011# Springer-Verlag 2011

Abstract In all vertebrate animals, CD8+ cytotoxic Tlymphocytes (CTLs) are controlled by major histocompat-ibility complex class I (MHC-I) molecules. These arehighly polymorphic peptide receptors selecting and pre-senting endogenously derived epitopes to circulating CTLs.The polymorphism of the MHC effectively individualizesthe immune response of each member of the species. Wehave recently developed efficient methods to generaterecombinant human MHC-I (also known as human leuko-cyte antigen class I, HLA-I) molecules, accompanyingpeptide-binding assays and predictors, and HLA tetramersfor specific CTL staining and manipulation. This hasenabled a complete mapping of all HLA-I specificities(“the Human MHC Project”). Here, we demonstrate thatthese approaches can be applied to other species. Wesystematically transferred domains of the frequentlyexpressed swine MHC-I molecule, SLA-1*0401, onto a

HLA-I molecule (HLA-A*11:01), thereby generating re-combinant human/swine chimeric MHC-I molecules aswell as the intact SLA-1*0401 molecule. Biochemicalpeptide-binding assays and positional scanning combinato-rial peptide libraries were used to analyze the peptide-binding motifs of these molecules. A pan-specific predictorof peptide–MHC-I binding, NetMHCpan, which wasoriginally developed to cover the binding specificities ofall known HLA-I molecules, was successfully used topredict the specificities of the SLA-1*0401 molecule aswell as the porcine/human chimeric MHC-I molecules.These data indicate that it is possible to extend thebiochemical and bioinformatics tools of the Human MHCProject to other vertebrate species.

Keywords Recombinant MHC . Peptide specificity .

Binding predictions

Introduction

Major histocompatibility complex class I (MHC-I) mole-cules are found in all vertebrate animals where they play acrucial role in generating specific cellular immuneresponses against viruses and other intracellular pathogens.They are highly polymorphic proteins that bind 8–11 aminoacid long peptides derived from the intracellular proteinmetabolism. The resulting heterotrimeric complexes—con-sisting of the MHC-I heavy chain, the monomorphic lightchain, beta-2 microglobulin (β2m), and specifically boundpeptides—are translocated to the cell surface where theydisplayed as target structures for peptide-specific, MHC-I-restricted CTLs. If a peptide of foreign origin is detected,the T cells may become activated and kill the infected targetcell.

Electronic supplementary material The online version of this article(doi:10.1007/s00251-011-0555-3) contains supplementary material,which is available to authorized users.

L. E. Pedersen :M. Harndahl :M. Rasmussen :K. Lamberth :S. Buus (*)Laboratory of Experimental Immunology,Faculty of Health Sciences, University of Copenhagen,Panum 18.3.12, Blegdamsvej 3B,2200 Copenhagen, Denmarke-mail: [email protected]

W. T. GoldePlum Island Animal Disease Center,Agricultural Research Service, USDA,Greenport, NY, USA

O. Lund :M. NielsenCenter for Biological Sequence Analysis,Technical University of Denmark,Copenhagen, Denmark

Immunogenetics (2011) 63:821–834DOI 10.1007/s00251-011-0555-3

http://dx.doi.org/10.1007/s00251-011-0555-3

MHC-I is extremely polymorphic. In humans, more than3,400 different human leukocyte antigen class I (HLA-I)molecules have been registered (as of January 2011), andthis number is currently growing rapidly as more efficientHLA typing techniques are employed worldwide. Thepolymorphism of the MHC-I molecule is concentrated inand around the peptide-binding groove, where it determinesthe peptide-binding specificity. Due to this polymorphism,it is highly unlikely that any two individuals will share thesame set of HLA-I molecules thereby presenting the samepeptides and generating T cell responses of the samespecificities—something, that otherwise would give micro-organisms a strong evolutionary chance of escape. Rather,this polymorphism can be seen as diversifying peptidepresentation thereby individualizing T cell responses andreducing the risk that escape variants of microorganismsmight evolve.

In 1999, we proposed that all human MHC specificitiesshould be mapped (“the Human MHC Project”) as apreamble for the application of MHC information andtechnologies in humans (Buus 1999). Since then, we havedeveloped large-scale tools that are generally applicabletowards this goal: production, analysis, prediction andvalidation of peptide–MHC interactions (Ferre et al. 2003;Harndahl et al. 2009; Hoof et al. 2009; Larsen et al. 2005;Lundegaard et al. 2008; Nielsen et al. 2003, 2007;Ostergaard et al. 2001; Pedersen et al. 1995; Stranzl et al.2010; Stryhn et al. 1996), and a “one-pot, read-and-mix”HLA-I tetramer technology for specific T cell analysis(Leisner et al. 2008). Here, we demonstrate that many ofthese tools can be transferred to other vertebrate animals asexemplified by an important livestock animal, the pig. Wehave successfully generated a recombinant swine leukocyteantigen I (SLA-I) protein, SLA-1*0401, one of the mostcommon SLA molecules of swine (Smith et al. 2005).Using this protein, we have developed the accompanyingbiochemical peptide-binding assays and demonstrated thatthe immunoinformatics tools originally developed to coverall HLA-I molecules, despite the evolutionary distance, canbe applied to SLA-I molecules. We suggest that the “humanMHC project” can be extended to cover other species ofinterest.

Materials and methods

Peptides and peptide libraries

All peptides were purchased from Schafer-N, Denmark(www.schafer-n.com). Briefly, they were synthesized bystandard 9-fluorenylmethyloxycarbonyl (Fmoc) chemistry,purified by reversed-phase high-performance liquid chro-matography (to at least >80% purity, frequently 95–99%

purity), validated by mass spectrometry, and quantitated byweight.

Positional scanning combinatorial peptide libraries(PSCPL) peptides were synthesized using standard solid-phase Fmoc chemistry on 2-chlorotrityl chloride resins.Briefly, an equimolar mixture of 19 of the common Fmocamino acids (excluding cysteine) was prepared for eachsynthesis and used for coupling in 8 positions, whereas asingle type of Fmoc amino acid (including cysteine) wasused in one position. This position was changed in eachsynthesis starting with the N-terminus and ending with theC-terminus. In one synthesis, the amino acid pool was usedin all nine positions. The library therefore consisted of 20×9+1=181 individual peptide libraries:

& Twenty PSCPL sublibraries describing position 1: AX8,CX8, DX8, …. YX8

& Twenty PSCPL sublibraries describing position 2:XAX7, XCX7, XDX7, …. XYX7

& etc& Twenty PSCPL sublibraries describing position 9: X8A,

X8C, X8D, …. X8Y& A completely random peptide library: X9

X denotes the random incorporation of amino acids fromthe mixture, whereas the single letter amino acid abbrevi-ation is used to denote identity of the fixed amino acid.

The peptides in each synthesis were cleaved from theresin in trifluoroacetic acid/1,2-ethanedithiol/triisopropylsi-lane/water 95:2:1:3 v/v/v/v, precipitated in cold diethylether,and extracted with water before desalting on C18 columns,freeze drying, and weighting.

Recombinant constructs encoding chimericand SLA-1*0401 molecules

A synthetic gene encoding a transmembrane-truncatedfragment encompassing residues 1 to 275 of human HLA-A*11:01 alpha chain followed by a FXa–BSP–HAT tag(FXa = factor Xa cleavage site comprised of the amino acidsequence IEGR, BSP = biotinylation signal peptide, HAT =histidine affinity tag for purification purposes; see OnlineResource 1) had previously been generated and insertedinto the pET28 expression plasmid (Novagen) (Ferre et al.2003). Synthetic genes encoding the corresponding frag-ments of the SLA-1*0401 alpha chain (α1α2) and α3,respectively, (Sullivan et al. 1997) were purchased fromGenScript. To exchange domains and generate chimerichuman/swine MHC-I gene constructs, a type II restrictionendonuclease-based cloning strategy (SeamLess® Strate-gene; Cat#214400, Revision#021003a), with modifications,was used. All primers were purchased HPLC-purified fromEurofins MWG Operon (Ebersberg, Germany), and all PCRamplifications were performed in a DNA Engine Dyad

822 Immunogenetics (2011) 63:821–834

http://www.schafer-n.com

PCR instrument (MJ Research, MN, USA). All constructswere validated by DNA sequencing. The following MHC-Iheavy chain constructs were made HHH, HHP, HPP, PHP,and PPP, where the first, second, and third letter indicatesdomains α1 (positions 1–90), α2 (positions 91–181), and α3

(positions 182–275), respectively, and H indicates that thedomain is of HLA-A*11:01 origin, whereas P indicates thatit is of SLA-1*0401 origin.

Constructs were transformed into DH5α cells, cloned,and sequenced (ABI Prism 3100Avant, Applied Biosys-tems) (Hansen et al. 2001). Validated constructs of interestwere transformed into an Escherichia coli production cellline, BL21(DE3), containing the pACYC184 expressionplasmid (Avidity, Denver, USA) containing an isopropyl-β-d-1-thiogalactopyranoside (IPTG)-inducible BirA geneto express biotin-ligase. This leads to almost complete invivo biotinylation of the desired product (Leisner et al.2008).

Expression of recombinant proteins

To maintain the pET28-derived plasmids, the media wassupplemented with kanamycin (50 μg/ml) throughout theexpression cultures. When appropriate, the media wasfurther supplemented with chloroamphenicol (20 μg/ml)to maintain the BirA containing pACYC184 plasmid. E.coli BL21(DE3) cells transformed with appropriate plas-mids were grown for 5 h at 30°C, and a 10-ml sampleadjusted to OD(600)=1 was then transferred to a 2-l fed-batch fermentor (LabFors®). To induce protein expression,IPTG (1 mM) was added at OD(600)∼25 and the culture wascontinued for an additional 3 h at 42°C (for in vivobiotinylation of the product, the induction media wasfurther supplemented with biotin (Sigma #B4501, 125 μg/ml)). Samples were analyzed by reducing sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE)before and after IPTG induction. At the end of theinduction culture, protease inhibitor (PMSF, 80 μg/l) wasadded, and cells were lysed in a cell disrupter (ConstantCell Disruptor Systems set at 2,300 bar) and the releasedinclusion bodies were isolated by centrifugation (SorvalRC6, 20 min, 17,000×g). The inclusion bodies werewashed twice in PBS, 0.5% NP-40 (Sigma), and 0.1%deoxycholic acid (Sigma) and extracted into urea–Trisbuffer (8 M urea, 25 mM Tris, pH 8.0), and anycontaminating DNA was precipitated with streptomycinsulfate (1% w/v).

MHC class I heavy chain purification

The dissolved MHC-I proteins were purified by Ni2+/IDAmetal chelating affinity column chromatography followedby Q-Sepharose ion exchange column chromatography,

hydrophobic interaction chromatography, and eventually bySuperdex-200 size exclusion chromatography. Fractionscontaining MHC-I heavy chain molecules were identifiedby A280 absorbance and SDS-PAGE and pooled. Through-out purification and storage, the MHC-I heavy chainproteins were dissolved in 8 M urea to keep themdenatured. Note that the MHC-I heavy chain proteins atno time were exposed to reducing conditions. This allowedpurification of highly active pre-oxidized moieties aspreviously described (Ostergaard et al. 2001). Proteinconcentrations were determined by bicinchoninic acidassay. The degree of biotinylation (usually >95%) wasdetermined by a gel-shift assay (Leisner et al. 2008). Thepre-oxidized, denatured proteins were stored at −20°C inTris-buffered 8 M urea.

Recombinant constructs encoding human and porcinebeta-2 microglobulin

Recombinant human β2m was expressed and purified asdescribed elsewhere (Ostergaard et al. 2001), (Ferre et al.2003). Using a previously reported E. coli codon-optimized gene encoding human β2m as template(Pedersen et al. 1995), a gene encoding porcine β2mwas generated by multiple rounds of site-directedmutagenesis (QuikChange® Stratagene, according to themanufacturer’s instructions) (Online Resource 2). Briefly,the genes encoding human or pig β2m were N-terminallyfused to a histidine affinity encoding tag (HAT) followedby a restriction enzyme encoding tag (FXa), inserted intothe pET28 vector and expressed in inclusion bodies in E.coli. The fusion proteins were extracted into 8 M urea,purified by immobilized metal affinity chromatography(IMAC), and refolded by dilution. The fusion tags werethen removed by FXa restriction protease digestion. Theliberated intact and native human or pig β2m werepurified by IMAC and gel filtration chromatography,analyzed by SDS-PAGE analysis, concentrated, andstored at −20°C until use (Fig. 1).

Purification and refolding of recombinant porcine β2mproteins

Porcine β2m was purified in the same way as human β2m(Ostergaard et al. 2001; Ferre et al. 2003). Briefly, the urea-dissolved β2m protein was purified by Ni2+/IDA metalchelating affinity column chromatography, refolded bydrop-wise dilution into an excess refolding buffer understirring (25 mM Tris, 300 mM urea, pH 8.00), and thenconcentrated (VivaFlow, 10 kDa). The refolded product waspurified by Ni2+/IDA metal chelating affinity columnchromatography again (this time in aqueous buffer, i.e.,without urea). Fractions containing HAT-pβ2m were iden-

Immunogenetics (2011) 63:821–834 823

tified by SDS-PAGE and pooled. Removal of the HAT tagwas performed by cleavage with factor Xa restrictionprotease (FXa) followed by renewed purified by Ni2+/IDA metal chelating affinity and Superdex200 gel filtrationcolumn chromatography, concentrated by spin ultrafiltration(10 kDa), mixed 1:1 with glycerol, and stored at −20°C.

SDS-PAGE analysis

Protein samples were mixed 1:1 in SDS sample buffer (4%SDS, 17.4% glycerol, 0.003% bromophenol blue, 0.125 MTris, 8 mM IAA (iodoacetamide)) with or without reducingagent (2-mercaptoethanol) as indicated, boiled for 3 min,spun at 20,000×g for 1 min, and loaded onto a 12% or 15%running gel with a 5% stacking gel. Gels were run at 180 V,40 mA for 50 min.

Peptide–MHC class I interaction measured by radioassayand spun column chromatography

A HLA-A*11:01-binding peptide, KVFPYALINK (non-natural consensus sequence A3CON1 (Sette et al. 1994)),was radiolabeled with iodine (125I) using a chloramine-Tprocedure (Hunter and Greenwood 1962). Dose titrationsof MHC-I heavy chains (HHH or HHP) were diluted intorefolding buffer (Tris–maleate–PBS) and mixed with β2m(human or porcine) and radiolabeled peptide, and incubat-ed at 18°C overnight. Then binding of radiolabeledpeptide to MHC-I was determined in duplicate bySephadex™ G50 spun column gel chromatography aspreviously described (Buus et al. 1995). MHC boundpeptide eluted in the excluded volume, whereas freepeptide was retained on the microcolumn. Both fractions

were counted by gamma spectroscopy, and the fractionpeptide bound was calculated as excluded radioactivitydivided by total radioactivity.

To examine the affinity of the interaction, increasingconcentrations of unlabeled competitor peptide were added.When conducted under limiting concentrations of MHC-Imolecule, the concentration of competitor peptide needed toeffect 50% inhibition of the interaction, the IC50, is anapproximation of the affinity of the interaction betweenMHC-I and the competitor peptide.

Peptide–MHC class I interaction measuredby an enzyme-linked immunosorbent assay

Peptide–MHC-I interaction was also measured in a modi-fied version of a previously described enzyme-linkedimmunosorbent assay (ELISA) (Sylvester-Hvid et al.2002). Briefly, denatured biotinylated recombinant MHC-Iheavy chains were diluted into a renaturation buffercontaining β2m and graded concentrations of the peptideto be tested and incubated at 18°C for 48 h allowingequilibrium to be reached. We have previously demonstrat-ed that denatured MHC molecules can de novo foldefficiently, however, only in the presence of appropriatepeptide. The concentration of peptide–MHC complexesgenerated was measured in a quantitative sandwich ELISA(using streptavidin as capture layer and the monoclonalanti-β2m antibody, BBM1, as detection layer) and plottedagainst the concentration of peptide offered (Sylvester-Hvidet al. 2002). A prefolded, biotinylated FLPSDYFPSV/HLA-A*02:01 (Kast et al. 1994) complex was used asstandard. Because the effective concentration of MHC (3–5 nM) used in these assays is below the equilibriumdissociation constant (KD) of most high-affinity peptide–MHC interactions, the peptide concentration, ED50, leadingto half-saturation of the MHC is a reasonable approxima-tion of the affinity of the interaction.

Using peptide libraries to perform an unbiasedanalysis of MHC specificity

The experimental strategy of PSCPL has previously beendescribed (Stryhn et al. 1996). The construction of thesublibraries and the ELISA-driven quantitative measure-ments of MHC interaction are as given above.

Briefly, the relative binding (RB) affinity of each PSCPLsublibrary was determined as RB (PSCPL)=ED50 (X9)/ED50 (PSCPL) (where ED50 is the concentration needed tohalf-saturate a low concentration of MHC-I molecules) andnormalized so that the sum of the RB values of the 20naturally occurring amino acids equals 20 (since peptideswith a given amino acid in a given position are 20 timesmore frequent in the corresponding PSCPL sublibrary than

Fig. 1 SDS-PAGE analysis of peak fractions from size exclusionchromatography of porcine β2m (11.4 kDa) after removal of histidineaffinity tag (HAT) by FXa protease digestion. Samples were mixed 1:1in non-reducing SDS sample buffer and loaded onto a 15%polyacrylamide gel

824 Immunogenetics (2011) 63:821–834

in the completely random X9 library). A RB value above 2was considered as the corresponding position and aminoacid being favored, whereas a RB value below 0.5 wasconsidered as being unfavorable (these thresholds representthe 95% confidence intervals). An anchor position (AP)value was calculated by the equation ∑(RB−1)2. A primaryanchor position is characterized by one or few amino acidsbeing strongly preferred and many amino acids beingunacceptable. We have arbitrarily defined anchor residuesas having an AP value above 15 (Lamberth et al. 2008).The peptide–SLA-I*0401 binding activity of each subli-brary was determined using previously published biochem-ical binding assay (ELISA) (Sylvester-Hvid et al. 2002)(with the modifications described above).

Sequence logos

Sequences logos describing the predicted binding motiffor each MHC molecule were calculated as described byRapin et al. (2010). In short, the binding affinity for a setof 1,000,000 random natural 9mer peptides was predictedusing the NetMHCpan method, and the 1% strongestbinding peptides were selected for construction of aposition-specific scoring matrix (PSSM). The PSSM wasconstructed as previously described (Nielsen et al. 2004)including pseudo-count correction for low counts. Next,sequence logos were generated from the amino acidfrequencies identified in the PSSM construction. For eachposition, the frequency of all 20 amino acids is displayedas a stack of letters. The total height of the stackrepresents the sequence conservation (the informationcontent), while the individual height of the symbolsrelates to the relative frequency of that particular symbolat that position. Letter shown upside-down are underrep-resented compared to the background (for details seeRapin et al. (2010)).

MHC distance trees

MHC distance trees were derived from correlations be-tween predicted binding affinities. For each allelic MHC-Imolecule, the binding affinity was predicted for a set of200,000 random natural peptides using the NetMHCpanmethod. Next, the distance between any two alleles wasdefined, as D=1−PCC, where PCC is the Pearsoncorrelation between the subset of peptides within thesuperset of top 10% best binding peptides for each allele.In this measure, two molecules that share a similar bindingspecificity will have a distance close to 0 whereas twomolecules with non-overlapping binding specificities wouldhave a distance close to 2. Using bootstrap, 100 suchdistance trees were generated, and branch bootstrap valuesand the consensus tree were calculated.

Results

Generation of chimeric MHC class I molecules

We have previously generated highly active, recombinanthuman MHC-I (HLA-I) molecules and accompanying high-throughput assays and bioinformatics prediction resources.Here, we transfer the underlying approaches to an importantdomesticated livestock animal, the pig, and its MHCsystem, the SLAs. MHC-I molecules are composed of aunique and highly variable distal peptide-binding platformconsisting of the alpha 1 (α1) and alpha 2 (α2) domains ofthe MHC-I heavy chain (HC) and a much more conservedproximal immunoglobulin-like membrane attaching stalkconsisting of the alpha 3 (α 3) domain of the HC non-covalently associated with the soluble MHC-I light chain(β2m).

A priori, the establishment of recombinant SLA mole-cules is complicated by the lack of validated reagents. Anyfailure could therefore be caused either by real technicalproblems in generating SLA molecules, or merely by a lackof information about strong peptide binders to the SLA inquestion. To reduce this uncertainty, we decided to migratefrom human to pig MHC-I in a step-wise manner andgenerate an intermediary chimeric MHC-I molecule com-posed of a well-known human peptide-binding platformattached to a SLA stalk, which might allow us to assesswhether we could generate a functional SLA stalk consist-ing of SLA-1*0401 α3 HC and pig β2m. To this end, weused the α1α2 domains of the HLA-A*11:01 molecule,which we expected should be able to bind a known high-affinity HLA-A*11:01-binding peptide (KVFPYALINK).This peptide could be 125I radiolabeled and used in a veryrobust peptide-binding assay testing whether the humanstalk could be replaced with the corresponding SLA stalk.Once that had been successfully established, the entireSLA-1*0401 molecule would be constructed and tested.

We have previously expressed and purified the extracellularsegment spanning positions 1–275 of the human HLA-A*11:01 in a denatured and pre-oxidized version that rapidlyrefold and bind appropriate target peptides (Ostergaard et al.2001; Ferre et al. 2003). Codon-optimized genes encodingthe corresponding segments of SLA-1*0401 (α1α2) andSLA-1*0401 (α3) were constructed as described in the“Materials and methods” section and used to replace theHLA-A*11:01 gene segment in the above construct gener-ating a new construct allowing for the expression of SLA-1*0401. For the generation of HLA-A*11:01/SLA-1*0401 chimeras, the genes encoding α1 (spanningpositions 1–90), α2 (spanning positions 91–181), and/orα3 (spanning positions 182–275) domains of HLA-A*11:01 and SLA-1*0401 were exchanged using Seam-Less and touch-down cloning strategies. Genes encoding

Immunogenetics (2011) 63:821–834 825

the extracellular segments 1–275 of the above natural orchimeric MHC-I molecules were C-terminally fused to abiotinylation tag (as indicated for SLA-1*0401 in OnlineResource 1), inserted into pET28, and expressed ininclusion bodies in E. coli (Fig. 2 shows SDS-PAGE oflysates of recombinant E. coli before and 3 h after IPTGinduction). The fusion proteins were extracted into 8 M urea(without any reducing agents), purified by ion exchange,hydrophobic and gel filtration chromatography (all conducted

in 8M urea, without any reducing agents) (Fig. 3 shows SDS-PAGE of the purified SLA-1*0401 after gel filtration),concentrated, and stored in urea at −20°C.

Testing a chimeric molecule consisting of a SLA-1*0401stalk and a HLA-A*11:01 peptide-bindingplatform—comparing human versus porcine β2m

To test the proximal immunoglobulin-like membraneattaching SLA stalk, we generated recombinant porcineβ2m and a chimeric human/porcine MHC-I heavy chainmolecule where the α1α2 were derived from the humanHLA-A*11:01, and the α3 was derived from the porcineSLA-1*0401. Since this construct contains the entirepeptide-binding platform of HLA-A*11:01, we reasonedthat the binding of the HLA-A*11:01 restricted peptide,KVFPYALINK, could be used as a functional readout ofthe refolding, activity, and assembly of the entire chimericmolecule including the porcine SLA stalk. For comparison,we tested the supportive capacity of human β2m andfolding ability of the entirely human HLA-A*11:01. A totalof four combinations could therefore be tested: porcine orhuman β2m in combination with either HHP or HHH(where the first letter indicates the origin of the α1 domain(Human HLA-A*11:01 or Porcine SLA-1*0401), thesecond letter the origin of the α2 domain, and the thirdletter the origin of the α3 domain). A concentration–titration of heavy chain was added to a fixed excessconcentration (3 μM) of β2m and a fixed trace concentra-tion (23 nM) of radiolabeled peptide. As shown in Figs. 4and 5, the four combinations gave almost the same heavychain dose titration with a half-saturation occurring around1–2 nM heavy chain. Porcine β2m supported folding of thechimeric (HHP) α chain slightly better than it supportedfolding of the human (HHH) α chain. Human β2m

Fig. 2 SDS-PAGE analysis of cell lysates of SLA-1*0401 expressionin E. coli before and after induction with IPTG. Samples were mixed1:1 in a reducing SDS sample buffer and loaded onto a 12%polyacrylamide gel

Fig. 3 SDS-PAGE analysis ofsize exclusion chromatographypeak fractions from SLA-1*0401 purification. Fractionnumbers are shown above eachlane. Samples were mixed 1:1 innon-reducing SDS sample bufferand loaded onto a 12%polyacrylamide gel. Lanes wereloaded with different volumes,indicated below each lane, toavoid overloading. An arrowindicates the band representingpurified SLA-1*0401-BSP-HATheavy chain (36,675 Da)

826 Immunogenetics (2011) 63:821–834

supported folding of HHP and HHH equally well. Thus, arecombinant SLA stalk can fold and support peptidebinding of the peptide-binding platform. These results alsosuggest that human β2m can support folding and peptidebinding of porcine MHC-I heavy chain molecules.

Using a positional scanning combinatorial peptide libraryapproach to perform an unbiased analysis of the specificityof SLA-1*0401 and human–pig chimeric MHC classI molecules

Using human β2m to support folding, the recombinant SLA-1*0401 and human–pig chimeric MHC-I molecule weretested for peptide binding. We have previously describedhow PSCPL can be used to perform an unbiased analysis ofMHC-I molecules (Stryhn et al. 1996). A PSCPL consists of20 sublibraries for each position where one of each of the 20natural amino acids have been locked and all other positionscontain random amino acids. Analyzing how much of each

PSCPL sublibrary is needed to support MHC-I folding (seeexamples in Fig. 6) and comparing each sublibrary with acompletely random library, the effect of any amino acid inany position can be examined and expressed as a RB value.Further, an AP value calculated as the sum of squareddeviations of RB values for each position can be used toidentify the most prominent anchor position (see “Materialsand methods” for calculations). Thus, the specificity of anonamer binding MHC-I molecule can be analyzed compre-hensively with 9×20+1 completely random library=181sublibraries (Stryhn et al. 1996).

Here, this approach was used to perform a completeexperimental analysis of SLA-1*0401 and a limitedanalysis of the chimeric HPP and PHP molecules. Anonamer PSCPL analysis of SLA-1*0401 can be seen inTable 1. AP values identified positions 9, 3, and 2 (in thatorder of importance) as the anchor positions of SLA-1*0401. In position 9, the amino acid preferences weredominated by the large and bulky aromatic tyrosine (Y),tryptophane (W), and phenylalanine (F), all having RBvalues above 4 (Table 1). In the almost equally importantposition 3, preferences for negatively charged amino acidsglutamic acid (E) and aspartic acid (D) were observed. Inthe lesser important position 2, the most preferred aminoacids were the hydrophobic amino acids valine (V),isoleucine (I), and leucine (L), followed by the polar aminoacids threonine (T) and serine (S).

Finally, a very limited PSCPL analysis was performed forthe two chimeric human HLA-A*11:01/porcine SLA-1*0401MHC-I molecules, HPP and PHP (Table 2). For both chimericmolecules, it could be demonstrated that position 9 is astrong anchor position. The positively charged amino acids,arginine (R) and lysine (K), were preferred in position 9 ofthe chimeric HPP molecule, whereas the aromatic aminoacid, tyrosine (Y), was exclusively preferred in position 9 ofthe chimeric PHP molecule.

-2 -1 0 10

20

40

60

80Porcine β2m

Human β2m

log(HHP concentration,nM)

125 I-

pep

tid

e,B

ou

nd

fra

ctio

n (

%)

Fig. 5 MHC-I complex formation of the chimeric class I moleculeHHP (HLA-A*11:01 (α1α2), SLA-1*0401 (α3)) with a known HLA-A*11:01-binding peptide (KVFPYALINK), and human (open squares)versus porcine β2m (filled circles). The affinity of HHP to porcine andhuman β2m was determined as 0.774 nM (95% confidence interval,0.630 to 0.951) and 1.19 nM (95% confidence interval, 0.958 to 1.47),respectively

-2 -1 0 10

20

40

60

80Porcine β2m

Human β2m

log(HLA-A*11:01 concentration,nM)

125 I-

pep

tid

e,B

ou

nd

fra

ctio

n (

%)

Fig. 4 HLA-A*11:01 complex formation with a known HLA-A*11:01-binding peptide (KVFPYALINK) using either human (opensquares) or porcine (filled circles) β2m. The affinity of HLA-A*11:01to porcine and human β2m was determined as 2.11 nM (95%confidence interval, 1.96 to 2.27) and 1.07 nM (95% confidenceinterval, 0.871 to 1.32), respectively

-5 -4 -3 -2 -1 0 1 20.00

0.25

0.50

0.75

1.00

1.25

Serine

Threonine

Valine

Tryptophan

Tyrosine

X9

log(peptide, µM)

nM

pM

HC

-I c

om

ple

x

Fig. 6 PSCPL position 9 sublibrary analysis of the SLA-1*0401peptide-binding motif. The amino acids valine, serine, and threonineare disfavored in position 9, evident by a decrease in affinitycompared to the reference peptide (X9), whereas the large and bulkyamino acids tryptophan and tyrosine are favored as seen by anincrease in affinity compared to the reference peptide

Immunogenetics (2011) 63:821–834 827

The positively charged amino acids, arginine (R) and lysine(K), were preferred in position 9 of the chimeric HPP moleculesimilar to the position 9 specificity of the HLA-A*11:01molecule. In contrast, the aromatic amino acid tyrosine (Y) waspreferred in position 9 of the chimeric PHP molecule similar tothe position 9 specificity of the SLA-1*0401 molecule.

Using NetMHCpan to predict peptides that bindto SLA-1*0401 or to human–pig chimeric MHCclass I molecules

Our recently described neural network-driven bioinformaticspredictor, NetMHCpan (version 2.0), has been trained onabout 88,000 peptide-binding data points representing morethan 80 different MHC-I molecules (primarily HLA-A andHLA-B molecules). We have previously shown thatNetMHCpan is an efficient tool to identify peptides thatbind to HLA molecules where no prior data exist (Nielsen etal. 2007) and demonstrated that NetMHCpan can beextended to MHC-I molecules of other species1 (Hoof etal. 2009). We applied NetMHCpan to our peptide repositoryof about 10,000 peptides, which over the past decade have

been selected to scan infectious agents (e.g., SARS andinfluenza, Sylvester-Hvid et al. 2004; Wang et al. 2010),improve coverage of MHC-I specificities (e.g., Buus et al.2003; Christensen et al. 2003), etc. We extracted 29 peptidesas predicted binders to either the SLA-1*0401, the HPP, orthe PHP human/porcine chimeric class I molecules (some ofthe peptides were predicted to bind to two or even all threeof these molecules). All these peptide–MHC-I combinationswere tested for binding (Table 3); 13 of 14 peptides bound tothe SLA-1*0401 molecule with an affinity (IC50 value) betterthan 500 nM (6 with an affinity less than 50 nM); all 13peptides tested on the PHP molecule were strong binderswith IC50 values below 50 nM; and 3 of 12 peptides testedon the HPP molecule bound with an affinity better than500 nM. Of the 39 peptide–MHC-I combinations tested, 20(51%) were found to be good binders, 9 (23%) were averagebinders, and 10 (26%) did not bind well (Table 3). This is instark contrast to the 0.5% frequency of binders amongrandomly selected peptides (Yewdell and Bennink 1999).

Next, the NetMHCpan method was used to generatePSSMs and sequence logos from the corresponding aminoacid frequencies as described by Nielsen et al. (2004). Foreach position, the frequencies of all 20 amino acids weredisplayed as a stack of letters showing the sequenceconservation/information content (the height of the entire

1 A preliminary report of SLA-1*0401 binding was given in Hoof etal. (2009).

SLA-1*0401 Amino acid position in peptide

1 2 3 4 5 6 7 8 9

A 1.7 0.8 1.6 1.1 0.8 0.4 1.7 0.7 0.0

C 0.8 0.2 0.4 2.7 1.1 1.0 0.7 1.7 0.0

D 0.4 0.7 7.7 2.1 0.8 0.1 0.5 1.8 0.0

E 1.6 0.0 2.8 0.6 1.0 1.4 0.9 0.9 0.2

F 0.4 0.2 0.8 0.0 1.0 1.1 1.1 1.1 5.8

G 0.5 0.3 0.1 1.5 1.1 1.4 0.0 1.0 0.0

H 0.5 0.2 0.1 0.1 1.3 2.4 0.5 0.7 0.4

I 0.7 3.4 0.3 0.8 0.8 0.9 0.7 0.9 0.3

K 1.1 0.2 0.1 0.4 0.4 1.2 0.4 0.9 0.0

L 0.6 2.1 0.2 0.6 1.0 1.7 1.0 1.2 0.6

M 1.7 1.9 1.0 1.2 1.4 1.4 1.2 1.4 0.7

N 1.4 0.5 0.6 1.4 1.9 0.7 1.0 1.9 0.0

P 0.0 0.0 0.1 1.6 0.5 1.1 1.9 1.2 1.1

Q 1.0 0.9 0.2 1.0 0.5 0.4 0.8 0.7 0.0

R 2.3 0.0 0.1 0.6 2.4 0.9 1.0 0.5 0.2

S 1.8 2.4 1.8 1.1 0.8 1.5 1.6 0.7 0.0

T 1.4 2.5 1.6 0.7 0.9 0.5 2.0 0.7 0.0

V 1.1 3.7 0.3 0.5 0.7 0.8 0.9 0.6 0.8

W 0.3 0.0 0.2 0.6 1.0 0.8 0.6 1.4 5.6

Y 0.6 0.0 0.5 1.3 0.6 0.2 1.6 0.3 4.3

Sum 20 20 20 20 20 20 20 20 20

AP value 7 28 57 8 4 6 5 4 67

Table 1 Binding motif of theSLA-1*0401 allele determinedby PSCPL strategy

The normalized relative binding(RB) value indicates whether anamino acid is favored (RB>2,bold numbers) or disfavored(RB<0.5, italic numbers) in agiven peptide position. Theanchor position (AP) value isgiven by the equation∑(RB−1)2 . The importantanchor positions 2, 3, and 9 forSLA-1*0401 are underlined

828 Immunogenetics (2011) 63:821–834

stack) and the relative frequency of amino acids (the heightof the individual amino acids). Figure 7 shows a specificitytree clustering of the SLA-1*0401 molecule compared toprevalent representatives of the 12 common HLA super-types that NetMHCpan originally intended to cover (Lundet al. 2004). By this token, SLA-1*0401 most closelyresembles that of HLA-A*01:01.

The limited PSCPL analysis of the chimeric MHC-Imolecules revealed strong P9 signals with specificities thatseemed to be determined by the origin of the α1 domain: theHPP chimera showed an HLA-A*11:01-like P9 specificity,whereas the PHP chimera showed a SLA-1*0401/HLA-A*01:01-like specificity. Since the NetMHCpan predictorsuccessfully captured these chimeric specificities (see above),we reasoned that the predictor might also be used to performin silico dissection of these specificities and used the P9specificity as an example of such an in silico analysis. TheNetMHCpan predictor considers a pseudo-sequence consist-ing of 34 polymorphic positions, which contain residues thatare within 4.0 Å of the atoms of bound nonamer peptides(Nielsen et al. 2007). Of the 34 positions of the pseudo-sequence, 10 delineates the P9 binding pocket; however, only3 of these, positions 74, 77, and 97, differ between SLA-1*0401 and HLA-A*11:01. To explore the effect of these

three residues, we performed in silico experiments where weexamined single substitutions Y74D, G77D, and S97I (theletter before the position number indicates the SLA-A*0401single letter residue, whereas the letter after indicates theHLA-A*11:01 residue) as well as the corresponding triplesubstitution (YGS-DDI). As described above, PSSMs weregenerated for each of the in silico molecules followed by aspecificity tree clustering (including SLA-A*0401, HLA-A*01:01, and HLA-A*11:01). Figure 8 shows this tree alongwith the sequence logo plots showing the predicted bindingspecificity of each in silico MHC-I molecule. Albeit theY74D and G77D single substitutions showing some of thepositively charged P9 peptide residue preference of HLA-A*11:01, they still clustered with HLA-A*01:01. In contrast,the in silico (YGS-DDI) triple substitution clustered with theHLA-A*11:01. This suggests that the NetMHCpan method iscapable of defining the residues of the F pocket thatdetermine the specificity of position 9.

Discussion

We have previously suggested that the specificities of theentire human MHC-I system should be solved (“the human

Molecule: HPPPosition in peptide

Molecule: PPPPosition in peptide

Molecule: PHPPosition in peptide

Aminoacid

2 9 Aminoacid

2 9 Aminoacid

2 9

A 0.3 0.4 A 0.8 0.0 A 1.0 0.0

C 0.2 0.4 C 0.2 0.0 C 0.0 0.0

D 0.2 0.4 D 0.7 0.0 D 0.0 0.0

E 0.2 0.4 E 0.0 0.2 E 0.0 0.0

F 0.0 0.4 F 0.2 5.8 F 0.1 0.5

G 0.2 0.4 G 0.3 0.0 G 0.4 0.0

H 0.2 0.4 H 0.2 0.4 H 0.2 0.5

I 2.4 0.4 I 3.4 0.3 I 2.7 0.1

K 0.2 2.3 K 0.2 0.0 K 0.1 0.0

L 0.0 0.4 L 2.1 0.6 L 1.8 0.4

M 5.3 0.4 M 1.9 0.7 M 2.1 0.7

N 0.0 0.4 N 0.5 0.0 N 0.2 0.0

P 0.2 0.4 P 0.0 1.1 P 0.0 0.7

Q 0.0 0.4 Q 0.9 0.0 Q 1.1 0.0

R 0.2 9.6 R 0.0 0.2 R 0.2 0.0

S 3.4 0.4 S 2.4 0.0 S 2.8 0.0

T 4.8 0.4 T 2.5 0.0 T 3.6 0.0

V 1.6 0.4 V 3.7 0.8 V 3.3 0.3

W 0.2 0.4 W 0.0 5.6 W 0.1 0.2

Y 0.2 0.4 Y 0.0 4.3 Y 0.2 16.5

Sum 20 20 Sum 20 20 Sum 20 20

AP value 51 81 AP value 28 67 AP value 30 254

Table 2 Comparison of PSCPLderived binding motifs for thethree MHC-I heavy chains HPP,PHP, and PPP (hα1pα2pα3,pα1hα2pα3, and SLA-I*0401)regarding peptide positions 2and 9

RB and AP values are definedas described in Table 1

Immunogenetics (2011) 63:821–834 829

MHC”, Buus 1999; Lauemoller et al. 2000). However, dueto the extreme polymorphism of the MHCs, any attempt toaddress the specificity of the entire MHC system is asignificant experimental undertaking. During the pastdecade, we have established a series of technologies tosupport a general solution of human MHC class I and IIspecificities. For MHC-I, this includes (1) a highly efficientE. coli expression system for production of recombinanthuman and mouse MHC-I molecules (both heavy chain andlight chain (β2m) molecules) (Pedersen et al. 1995;Ostergaard et al. 2001), (2) a purification system forobtaining the highly active pre-oxidized MHC-I heavychain species (Ferre et al. 2003), (3) a high-throughputhomogenous peptide–MHC-I binding assay for obtaininglarge data sets on peptide–MHC-I interactions (Harndahl etal. 2009), (4) a positional scanning combinatorial peptidelibrary approach for a robust and unbiased analysis of thespecificity of any MHC-I molecule (Stryhn et al. 1996), (5)an immunobioinformatics approach to generate predictorsof the peptide–MHC-I interaction, NetMHCpan, that allowspredictions to be made for any human MHC-I molecules,HLA-I, even those that have not yet been covered byexisting data set (Hoof et al. 2009; Nielsen et al. 2007), andfinally (6) we have demonstrated that pre-oxidized MHC-Imolecules can be used to generate MHC-I tetramers in asimple “one-pot, mix-and-read” manner (Leisner et al.2008). Here, we propose that the next goal should be toextend the overall approach to MHC-I molecules of otherspecies of interest. Mouse and rats have been extensivelystudied in the past, but much less reagents and informationhave accrued for the MHC-I molecules of other species.Here, we have used an important livestock animal, the pig,as a model system and demonstrated that it indeed ispossible to transfer the original human approach to otherspecies.

Before attempting to generate a recombinant version ofthe entire porcine SLA-1*0401 molecule, we grafted themore conserved membrane-proximal “stalk” (theimmunoglobulin-like class I heavy chain α3 and β2mdomains) of porcine SLA-1*0401 onto the peptide-bindingdomain of HLA-A*11:01 generating a chimeric human/porcine MHC-I molecule. This chimeric molecule retainedthe peptide-binding specificity of the HLA-A*11:01 mole-cule, and it clearly demonstrated that the recombinantporcine stalk was functional and, by inference, alsoproperly folded. It also suggests that the peptide-bindingspecificity of the distal domains do not crucially dependupon the identity of the proximal stalk. Further, comparingthe ability of human and porcine β2m to support MHC-Icomplex formation using either a human or a porcineMHC-I stalk, we demonstrated that every combination(porcine β2m/human-α3, porcine β2m/porcine-α3, humanβ2m/human-α3, and human β2m/porcine-α3) showed al-

Table 3 Peptide sequences and in vitro determined KD values for thethree different MHC molecules PPP (SLA-1*0401) (top), PHP(pα1hα2pα3) (middle), and HPP (hα1pα2pα3) (bottom), respectively

Peptide # Sequence Affinity (KD, nM)

PPP

1 HSNASTLLY 452

2 SSMNSFLLY 192

3 YSAEALLPY 202

4 MTFPVSLEY 15

5 MSSAAHLLY 103

6 STFATVLEY 32

7 MTAASYARY 16

8 HTSALSLGY 163

9 ASYQFQLPY 8

10 YANMWSLMY 386

11 STYQPLPLY 38

12 HTAAPWGSY Non-binder

13 ITMVNSLTY 169

14 ATAAATEAY 5

PHP

1 HSNASTLLY 5

2 SSMNSFLLY 2

4 MTFPVSLEY 1

5 MSSAAHLLY 1

6 STFATVLEY 3

9 ASYQFQLPY 1

13 ITMVNSLTY 2

15 STMPLSWMY 1

16 HMMAVTLFY 11

17 KTFEWGVFY 6

18 ATNNLGFMY 1

19 ATATWFQYY 3

20 ITMVNSLTY 2

HPP

2 SSMNSFLLY Non-binder

4 MTFPVSLEY Non-binder

6 STFATVLEY Non-binder

21 FSFGGFTFK Non-binder

22 KVFDKSLLY Non-binder

23 ATFSVPMEK Non-binder

24 RVMPVFAFK 17

25 AVYSSSMVK 127

26 AVARPFFAK 231

27 GSYFSGFYK Non-binder

28 LTFLHTLYK Non-binder

29 MTMRRRLFK Non-binder

The NetMHCpan percentage of success for predicting binders for thePPP, PHP, and HPP heavy chains were 93%, 100%, and 33%,respectively

830 Immunogenetics (2011) 63:821–834

most the same heavy chain dose titration with identicalhalf-saturations. These results illustrate the ability forporcine and human β2m to support complex formation ofSLA molecules and vice versa and suggest evolutionarythat the stalk is quite conserved.

Next, we generated the entire SLA-1*0401 heavy chain andsucceeded in generating complexes using human β2m as thelight chain and PCSPL as peptide donors. The latter solved thea priori problem of not knowing which peptides would beneeded to support proper folding of SLA-1*0401, and it did soin an unbiased manner. Furthermore, this approach is highlyefficient since it readily establishes a complete matrix

representing the amino acid preference for each amino acidand each position of a nonamer peptide. The specificity ofSLA-1*0401 shows two primary anchors: one in positions 9with a preference for aromatic amino acids and another inposition 3 with a preference for negatively charged aminoacids. In addition, the SLA-1*0401 features a secondary anchorin position 2 with hydrophobic or polar amino acid preferences.

An alternative approach to solve the problem of identifyingpeptides that support folding of MHC-I molecules of so farunknown specificity is to use our recently developed pan-specific predictor, NetMHCpan. The successful use of thispredictor to initiate peptide-binding studies was recently

Fig. 7 Specificity tree clustering of the SLA-1*0401 moleculecompared to prevalent representatives of the 12 common HLAsupertypes (Lund et al. 2004). The distance between any two MHCmolecules and the consensus tree is calculated as described in“Materials and methods”. All branch points in the tree have bootstrapvalues of 100%. Sequence logos of the predicted binding specificity

are shown for each molecule. In the logo, acidic amino acids [DE] areshown in red, basic amino acids [HKR] in blue, hydrophobic aminoacids [ACFILMPVW] in black, and neutral amino acids [GNQSTY]in green. The axis of the LOGOs indicates in all case positions onethrough nine of the motif, and the y-axis the information content (seeMaterials and methods)

Immunogenetics (2011) 63:821–834 831

demonstrated for HLA-A*3001 (Lamberth et al. 2008).Although originally developed to cover all HLA-A andHLA-B molecules, it has also been shown to extend toMHC-I molecules of other species (Hoof et al. 2009). Here,we demonstrate that the NetMHCpan predictor is capable ofextracting MHC-I sequence information across species andcorrectly relate this to peptide binding even in the absence ofany available data for the specific query MHC-I molecule,i.e., the SLA-1*0401 as well as the chimeric HPP(hα1pα2pα3) and PHP (pα1hα2pα3) molecules. It is notclear why binding of the PHP chimera was more efficientlypredicted than binding of the HPP chimera. One couldspeculate that NetMHCpan has not captured the effect of thedifferent positions of the pseudo-sequence equally well andnot all positions and pockets (and by inference—not allchimeric molecules) are therefore predicted equally well.

Using the NetMHCpan predictor to cluster SLA-1*0401and representative molecules of 12 human HLA supertypesaccording to predicted peptide-binding specificities, the SLA-1*0401 specificity closely resembled that of HLA-A*01:01(IEDB, http://www.immuneepitope.org/MHCalleleId/142,

accessed March 9th 2011). This result was also obviousfrom an inspection of the PSCPL analysis of the SLA-1*0401. The PSCPL analysis of the P9 specificity of theSLA-1*0401 and the two chimeric molecules suggestedthat the P9 specificity primarily was determined by theα1 domain. This contention was further strengthened by aNetMHCpan-driven in silico analysis of the residuesdelineating the F pocket, which interacts with P9. Thissuggests that NetMHCpan can be used to design andinterpret detailed experiments addressing the structure–function relationship of peptide–MHC-I interaction. In thecase of SLA-1*0401, NetMHCpan suggests that Y74,G77, and S97 play a prominent role in defining the P9 Fpocket. Whereas the NetMHCpan readily captured the P9anchor residue of SLA-1*0401, it did not capture the P3anchor (at least not in the 2.4 version). We surmise thatthis shortcoming is due to insufficient examples of the useof P3 anchors within the currently available peptide–MHC-I binding data. Inspecting the pseudo-sequence ofSLA-1*0401 and HLA-A*01:01 vs. HLA-A*11:01 sug-gests that the presence of an arginine in position 156 might

Fig. 8 Comparison of specific in silico mutations of the SLA-1*0401molecule and comparison with the two HLA molecules: HLA-A*11:01 and HLA-A*01:01. The distance between any two MHCmolecules and the consensus tree is calculated as described in“Materials and methods”. All branch points in the tree have bootstrapvalues of 100%. The SLA-1*0401 mutations are indicated as Y74D,G77D, and S97I, where the letter before the position number indicates

the SLA-1*0401 single letter residue and the letter after indicates theHLA-A*11:01 residue. YGS-DDI is the corresponding triple substi-tution. Sequence logos are calculated and visualized as described inFig. 7. The axis of the LOGOs indicates in all case positions onethrough nine of the motif, and the y-axis the information content (seeMaterials and methods)

832 Immunogenetics (2011) 63:821–834

http://www.immuneepitope.org/MHCalleleId/142

explain the preference for negatively charged amino acidresidues in P3. Future NetMHCpan-guided experimentscould pointedly address this question, and the resultingdata could complement existing data and be used to updateand improve the NetMHCpan predictor.

All in all the two complementary approaches, PSCPLand NetMHCpan, agreed on the specificity of the SLA-1*0401 molecule, as well as of the two chimeric MHC-Imolecules. Thus, the specificity of SLA-1*0401 appear tobe well established. This specificity has successfully beenused to search for foot-and-mouth disease virus (FMDV)-specific CTL epitopes in FMDV-vaccinated, SLA-1*0401-positive pigs, and the recombinant SLA-1*0401 moleculeshave been used to generate corresponding tetramers andstain pig CTLs (Patch et al. 2011). In conclusion, we herepresent a set of methods that can be used to generatefunctional recombinant MHC-I molecules, map theirspecificities and identify MHC-I-restricted epitopes, andeventually generate peptide–MHC-I tetramers for validationof CTL responses. This suite of methods is not onlyapplicable to humans, but potentially to any species ofinterest.

Acknowledgments We thank Lise Lotte Bruun Nielsen, AnneCaroline Schmiegelow, and Iben Sara Pedersen for their expertexperimental support. This work was in part supported by the DanishCouncil for Independent Research, Technology and ProductionSciences (274-09-0281) and by the National Institute of Health(NIH) (HHSN266200400025C).

References

Buus S (1999) Description and prediction of peptide-MHC binding:the 'human MHC project'. Curr Opin Immunol 11:209–213

Buus S, Stryhn A, Winther K, Kirkby N, Pedersen LO (1995) Receptor–ligand interactions measured by an improved spun columnchromatography technique. A high efficiency and high throughputsize separation method. Biochim Biophys Acta 1243:453–460

Buus S, Lauemoller SL, Worning P, Kesmir C, Frimurer T, Corbet S,Fomsgaard A, Hilden J, Holm A, Brunak S (2003) Sensitivequantitative predictions of peptide-MHC binding by a 'Query byCommittee' artificial neural network approach. Tissue Antigens62:378–384

Christensen JK, Lamberth K, Nielsen M, Lundegaard C, Worning P,Lauemoller SL, Buus S, Brunak S, Lund O (2003) Selectinginformative data for developing peptide–MHC binding predictorsusing a query by committee approach. Neural Comput 15:2931–2942

Ferre H, Ruffet E, Blicher T, Sylvester-Hvid C, Nielsen LL, HobleyTJ, Thomas OR, Buus S (2003) Purification of correctly oxidizedMHC class I heavy-chain molecules under denaturing conditions:a novel strategy exploiting disulfide assisted protein folding.Protein Sci 12:551–559

Hansen NJ, Pedersen LO, Stryhn A, Buus S (2001) High-throughputpolymerase chain reaction cleanup in microtiter format. AnalBiochem 296:149–151

Harndahl M, Justesen S, Lamberth K, Roder G, Nielsen M, Buus S(2009) Peptide binding to HLA class I molecules: homogenous,

high-throughput screening, and affinity assays. J Biomol Screen14:173–180

Hoof I, Peters B, Sidney J, Pedersen LE, Sette A, Lund O, Buus S,Nielsen M (2009) NetMHCpan, a method for MHC class Ibinding prediction beyond humans. Immunogenetics 61:1–13

Hunter W, Greenwood F (1962) Preparation of iodine-131 labelledhuman growth hormone of high specific activity. Nature194:495–496

Kast WM, Brandt RM, Sidney J, Drijfhout JW, Kubo RT, Grey HM,Melief CJ, Sette A (1994) Role of HLA-A motifs in identificationof potential CTL epitopes in human papillomavirus type 16 E6and E7 proteins. J Immunol 152:3904–3912

Lamberth K, Roder G, Harndahl M, Nielsen M, Lundegaard C,Schafer-Nielsen C, Lund O, Buus S (2008) The peptide-bindingspecificity of HLA-A*3001 demonstrates membership of theHLA-A3 supertype. Immunogenetics 60:633–643

Larsen MV, Lundegaard C, Lamberth K, Buus S, Brunak S, Lund O,Nielsen M (2005) An integrative approach to CTL epitopeprediction: a combined algorithm integrating MHC class Ibinding, TAP transport efficiency, and proteasomal cleavagepredictions. Eur J Immunol 35:2295–2303

Lauemoller SL, Kesmir C, Corbet SL, Fomsgaard A, Holm A,Claesson MH, Brunak S, Buus S (2000) Identifying cytotoxic Tcell epitopes from genomic and proteomic information: "thehuman MHC project". Rev Immunogenet 2:477–491

Leisner C, Loeth N, Lamberth K, Justesen S, Sylvester-Hvid C,Schmidt EG, Claesson M, Buus S, Stryhn A (2008) One-pot,mix-and-read peptide–MHC tetramers. PLoS One 3:e1678

Lund O, Nielsen M, Kesmir C, Petersen AG, Lundegaard C, Worning P,Sylvester-Hvid C, Lamberth K, Roder G, Justesen S, Buus S,Brunak S (2004) Definition of supertypes for HLA molecules usingclustering of specificity matrices. Immunogenetics 55:797–810

Lundegaard C, Lamberth K, Harndahl M, Buus S, Lund O, Nielsen M(2008) NetMHC-3.0: accurate web accessible predictions ofhuman, mouse and monkey MHC class I affinities for peptides oflength 8–11. Nucleic Acids Res 36:W509–W512

Nielsen M, Lundegaard C, Worning P, Lauemoller SL, Lamberth K,Buus S, Brunak S, Lund O (2003) Reliable prediction of T-cellepitopes using neural networks with novel sequence representa-tions. Protein Sci 12:1007–1017

Nielsen M, Lundegaard C, Worning P, Hvid CS, Lamberth K, Buus S,Brunak S, Lund O (2004) Improved prediction of MHC class Iand class II epitopes using a novel Gibbs sampling approach.Bioinformatics 20:1388–1397

Nielsen M, Lundegaard C, Blicher T, Lamberth K, Harndahl M, JustesenS, Roder G, Peters B, Sette A, Lund O, Buus S (2007) NetMHCpan,a method for quantitative predictions of peptide binding to anyHLA-A and -B locus protein of known sequence. PLoS One 2:e796

Ostergaard PL, Nissen MH, Hansen NJ, Nielsen LL, Lauenmoller SL,Blicher T, Nansen A, Sylvester-Hvid C, Thromsen AR, Buus S(2001) Efficient assembly of recombinant major histocompatibil-ity complex class I molecules with preformed disulfide bonds.Eur J Immunol 31:2986–2996

Patch JR, Pedersen LE, Toka FN, Moraes M, Grubman MJ, NielsenM, Jungersen G, Buus S, Golde WT (2011) Induction of foot-and-mouth disease virus-specific cytotoxic T cell killing byvaccination. Clin Vaccine Immunol 18:280–288

Pedersen LO, Stryhn A, Holter TL, Etzerodt M, Gerwien J, NissenMH, Thogersen HC, Buus S (1995) The interaction of beta 2-microglobulin (beta 2m) with mouse class I major histocompat-ibility antigens and its ability to support peptide binding. Acomparison of human and mouse beta 2m. Eur J Immunol25:1609–1616

Rapin N, Hoof I, Lund O, Nielsen M (2010) The MHC motif viewer:a visualization tool for MHC binding motifs. Curr. Protoc.Immunol. Chapter 18:Unit 18.17

Immunogenetics (2011) 63:821–834 833

Sette A, Sidney J, del Guercio MF, Southwood S, Ruppert J, DahlbergC, Grey HM, Kubo RT (1994) Peptide binding to the mostfrequent HLA-A class I alleles measured by quantitativemolecular binding assays. Mol Immunol 31:813–822

Smith DM, Lunney JK, Martens GW, Ando A, Lee JH, Ho CS,Schook L, Renard C, Chardon P (2005) Nomenclature forfactors of the SLA class-I system, 2004. Tissue Antigens65:136–149

Stranzl T, Larsen MV, Lundegaard C, Nielsen M (2010) NetCTLpan:pan-specific MHC class I pathway epitope predictions. Immuno-genetics 62:357–368

Stryhn A, Pedersen LO, Romme T, Holm CB, Holm A, Buus S (1996)Peptide binding specificity of major histocompatibility complexclass I resolved into an array of apparently independentsubspecificities: quantitation by peptide libraries and improvedprediction of binding. Eur J Immunol 26:1911–1918

Sullivan JA, Oettinger HF, Sachs DH, Edge AS (1997) Analysis ofpolymorphism in porcine MHC class I genes: alterations in

signals recognized by human cytotoxic lymphocytes. J Immunol159:2318–2326

Sylvester-Hvid C, Kristensen N, Blicher T, Ferre H, Lauemoller SL,Wolf XA, Lamberth K, Nissen MH, Pedersen LO, Buus S (2002)Establishment of a quantitative ELISA capable of determiningpeptide–MHC class I interaction. Tissue Antigens 59:251–258

Sylvester-Hvid C, Nielsen M, Lamberth K, Roder G, Justesen S,Lundegaard C, Worning P, Thomadsen H, Lund O, Brunak S,Buus S (2004) SARS CTL vaccine candidates; HLA supertype-,genome-wide scanning and biochemical validation. Tissue Anti-gens 63:395–400

Wang M, Larsen MV, Nielsen M, Harndahl M, Justesen S, DziegielMH, Buus S, Tang ST, Lund O, Claesson MH (2010) HLA class Ibinding 9mer peptides from influenza A virus induce CD4 T cellresponses. PLoS One 5:e10533

Yewdell JW, Bennink JR (1999) Immunodominance in majorhistocompatibility complex class I-restricted T lymphocyteresponses. Annu Rev Immunol 17:51–88

834 Immunogenetics (2011) 63:821–834

42

4. Paper II

Identification of peptides from foot-and-mouth disease virus structural proteins

bound by class I swine leukocyte antigen (SLA) alleles, SLA-1*0401 and SLA-2*0401

Pedersen LE, Harndahl M, Nielsen M, Patch JR, Jungersen G, Buus S, Golde WT.

Anim Genet. 2012 Sep 18. doi: 10.1111/j.1365-2052.2012.02400.x

This manuscript describes the determination of the complete SLA-2*0401 binding motif and

apply such PSCPL determined amino acid preferences, expressed by both SLA-1*0401 and SLA-

2*0401, to NetMHCpan based predictions. Followed by in vitro peptide binding analyses these

studies lead to the identification of several high-affinity binding peptides derived from FMDV.

The analyses and methods presented provide the characterization of potential T cell epitopes in

response to threatening diseases such as FMDV.

Identification of peptides from foot-and-mouth disease virusstructural proteins bound by class I swine leukocyte antigen (SLA)alleles, SLA-1*0401 and SLA-2*0401

L. E. Pedersen*†,‡, M. Harndahl†, M. Nielsen§, J. R. Patch*, G. Jungersen‡, S. Buus† and

W. T. Golde**Foreign Animal Disease Unit, Plum Island Animal Disease Center, Agricultural Research Service, USDA, Greenport, NY, USA. †Laboratory

of Experimental Immunology, Faculty of Health Sciences, University of Copenhagen, Copenhagen N., 2200, Denmark. ‡National Veterinary

Institute, Technical University of Denmark, Copenhagen, Denmark. §Center for Biological Sequence Analysis, Technical University of

Denmark, Copenhagen, Denmark.

Summary Characterization of the peptide-binding specificity of swine leukocyte antigen (SLA) class I

and II molecules is critical to the understanding of adaptive immune responses of swine

toward infectious pathogens. Here, we describe the complete binding motif of the SLA-

2*0401 molecule based on a positional scanning combinatorial peptide library approach.

By combining this binding motif with data achieved by applying the NetMHCpan peptide

prediction algorithm to both SLA-1*0401 and SLA-2*0401, we identified high-affinity

binding peptides. A total of 727 different 9mer and 726 different 10mer peptides within the

structural proteins of foot-and-mouth disease virus (FMDV), strain A24 were analyzed as

candidate T-cell epitopes. Peptides predicted by the NetMHCpan were tested in ELISA for

binding to the SLA-1*0401 and SLA-2*0401 major histocompatibility complex class I

proteins. Four of the 10 predicted FMDV peptides bound to SLA-2*0401, whereas five of

the nine predicted FMDV peptides bound to SLA-1*0401. These methods provide the

characterization of T-cell epitopes in response to pathogens in more detail. The development

of such approaches to analyze vaccine performance will contribute to a more accelerated

improvement of livestock vaccines by virtue of identifying and focusing analysis on bona

fide T-cell epitopes.

Keywords binding motif, prediction, SLA, T-cell epitope

Introduction

The major histocompatibility complex class I (MHC I)

molecules are highly polymorphic proteins that bind 8 to

11 amino acid (AA) peptides derived from the intracellular

protein metabolism. In humans, such MHC class I molecules

are referred to as human leukocyte antigens (HLA) and in

swine are termed swine leukocyte antigens (SLA). These

integral membrane proteins are found on all nucleated cells

in vertebrates and play a crucial role in the cell-mediated

immune response against viruses, other intracellular patho-

gens and transformed cells (cancer). When an endogenous

peptide engages in a productive interaction with a MHC

class I heavy chain and the non-polymorphic MHC

I-stabilizing subunit molecule, beta-2-microglobulin (b2m),

they generate a heterotrimeric peptide/MHC/b2m (pMHC)

complex that is displayed at the cell surface. When surface-

exposed MHC complexes incorporate foreign peptides,

derived from a virus for instance, these complexes serve as

targets for the T-cell receptor (TCR) on CD8+, MHC I-

restricted, peptide-specific, cytotoxic T lymphocytes (CTLs).

In this way, endogenous peptides are constantly presented

on the cell surface, signaling the cell’s current state of

integrity to the immune system of the host. If the peptide

presented is of foreign origin, T cells may become activated

and kill the infected or transformed cell (Doherty &

Zinkernagel 1975; Zinkernagel & Doherty 1975; Bevan

1995; Harty et al. 2000).

Address for correspondence

W.T. Golde, Plum Island Animal Disease Center, Agricultural Research

Service, USDA, PO Box 848, Greenport, NY 11944, USA.

E-mail: [email protected]

Accepted for publication 19 June 2012

doi: 10.1111/j.1365-2052.2012.02400.x

1Published 2012. This article is a U.S. Government work and is in the public domain in the USA.

The effect of polymorphism of the MHC I molecule is

reflected in the different peptide-binding motifs that different

class I molecules have, with peptide-binding anchor

positions (APs) typically in positions 1, 2, 3 and 9 (Sette

& Sidney 1999; Lamberth et al. 2008). More than 5400

different HLA-I alleles have been registered according to

the IMGT/HLA database (http://www.ebi.ac.uk/imgt/hla/

stats.html), and this number is growing rapidly as more

efficient HLA typing techniques are employed worldwide.

The same has been seen for SLA typing in recent years, with

highly specific primer-based typing systems being used with

success for both class I and class II genes in inbred as well as

outbred pigs (Ho et al. 2006, 2010a,b; Yeom et al. 2010).

As such, there is an equivalent need for biochemical data.

However, characterizing the peptide-binding motif for each

of these alleles is not feasible. Rather, application of the

NetMHCpan peptide-binding predictor algorithm (Nielsen

et al. 2007; Hoof et al. 2009) developed from the extensive

human database available provides a method for quantita-

tive predictions of peptide binding to any MHC protein.

Thus, potential peptides that will bind can be reduced from

hundreds (or even thousands) to a more dexterous number

that can be tested.

Previously, we reported that the bioinformatics tools

originally developed to cover all HLA-I molecules, despite

the evolutionary distance, can be applied to SLA-I molecules

(Hoof et al. 2009; Pedersen et al. 2011). Here, we have used

the NetMHCpan predictor to predict peptides likely to be

bound by the SLA-1*0401 and SLA-2*0401 molecules

using a NetMHCpan rank score threshold of 2.00%. In this

way, we were able to limit the number of possible peptides

derived from the structural proteins of foot-and-mouth

disease virus (FMDV), strain A24, from more than 700 to

<16. Furthermore, by comparing the sequences of these

peptides with the AA binding preferences exhibited by the

SLA molecule in question, we were able to further reduce

the number of peptides to be tested. The final peptide

candidates were then analyzed in vitro for the ability to

induce the formation of pMHC complexes with a high

enough affinity to allow proper folding using the recombi-

nant SLA-1*0401 and SLA-2*0401 class I heavy chains.

Through the production of recombinant SLA molecules,

mapping of their peptide-binding motifs using universal

peptide libraries, that is, the positional scanning combina-

torial peptide libraries (PSCPL) (Stryhn et al. 1996;

Lamberth et al. 2008), and the screening of AA sequences

from FMDV structural proteins for possible peptide/MHC

matches using NetMHCpan (Hoof et al. 2009), we were able

to identify several in vitro binding peptides and characterize

the SLA-2*0401 binding motif. Such data add to the

knowledge of the binding characteristics for SLA molecules

in general and could eventually lead to further insight into

specific viral-derived peptides that induce porcine CTLs

during infection. The result of taking these approaches is a

rapid and accurate capacity to identify T-cell epitopes and

the subsequent ability to measure specific T-cell immunity

following vaccination or infection (Patch et al. 2011).


Recombinant construct encoding the SLA-2*0401protein

SLA-2*0401 recombinant protein was produced using a

previously described standard expression system (Pedersen

et al. 2011). Briefly, a transmembrane, truncated fragment

encompassing residues 1–276 of the SLA-2*0401 alpha

chain (Genscript, Appendix S1) followed by an FXa-BSP-

HAT tag (FXa = factor Xa cleavage site comprising the

AA sequence IEGR = I (isoleucine), E (glutamine acid),

G (glycine), R (arginine), BSP = biotinylation signal peptide,

HAT = histidine affinity tag for purification purposes) was

inserted into a pET28a expression vector (Novagen). The

construct was transformed into DH5a cells and sequenced

(ABI Prism 3100Avant; Applied Biosystems) (Hansen et al.

2001). The validated construct of interest was transformed

into an Escherichia coli (E. coli) production cell line, BL21

(DE3), containing the pACYC184 expression plasmid (Avid-

ity) with an IPTG (Isopropyl-beta-D-thiogalactoside)-induc-

ible BirA gene to express biotin-ligase. This leads to almost

complete in vivo biotinylation of the desired product

(Schmidt et al. 2009).

Expression and purification of recombinant SLA-2*0401and human beta-2 microglobulin

SLA-2*0401 heavy chain proteins were produced as previ-

ously described (Pedersen et al. 2011). E. coli BL21(DE3) cells

containing the SLA proteins were lyzed in a cell disrupter

(Constant Cell Disruptor Systems set at 2300 bar), and the

released inclusion bodies were isolated by centrifugation

(Sorval RC6, 20 min, 17 000 g). The inclusion bodies were

washed twice in PBS, 0.5%NP-40 (Sigma), 0.1%deoxycholic

acid (DOC; Sigma) and extracted into urea–Tris buffer (8 M

urea, 25 mMTris, pH8.0). SLA-2*0401heavy chain proteinsextracted in 8 M ureawere purified as previously described for

the SLA-1*0401 molecule (Pedersen et al. 2011). In brief,

this was carried out by successive immobilized metal adsorp-

tion, hydrophobic interaction, and Superdex-200 size exclu-

sion chromatography. Throughout purification and storage,

the MHC I heavy chain proteins were dissolved in 8 M urea to

keep them denatured. The pre-oxidized, denatured proteins

were stored at�20 °C inTris-buffered 8 Murea. Recombinant

human b2m was expressed and purified as described else-

where (Ostergaard et al.2001; Ferre et al.2003). Complexing

the porcine alpha chains with human b2m and peptide

allowed the application of the ELISA for class I MHC folding

using mouse anti-human b2m, as described below and

previously (Pedersen et al. 2011). Reagents for porcine b2mare not available.

Published 2012. This article is a U.S. Government work and is in the public domain in the USA., doi: 10.1111/j.1365-2052.2012.02400.x

Pedersen et al.2


All peptides were purchased from Schafer-N. Briefly, they

were synthesized by standard 9-fluorenylmethyloxycarbonyl

(Fmoc) chemistry, purified by reverse-phase high-perfor-

mance liquid chromatography (to at least >80% purity,

frequently 95–99% purity), validated by mass spectrometry

and quantitated by weight.

PSCPL peptides were used to determine the SLA-2*0401-binding motif. The experimental strategy of PSCPL has

previously been described for both HLA (Stryhn et al. 1996)

and SLA (Pedersen et al. 2011) proteins. Briefly, pools of

random peptides with a single fixed AA in a fixed peptide

position for every AA in every peptide position were used to

determine the MHC-binding pocket motif of the MHC I

protein. PSCPL peptides were used in a quantitative ELISA

as described for the SLA-1*0401 molecule (Pedersen et al.

2011). The relative binding (RB) affinity of each PSCPL

sublibrary was determined as RB (PSCPL) = ED50(X9)/

ED50(PSCPL) and normalized, so that the sum of the RB

values of the 20 naturally occurring AA equaled 20 (ED50

being the concentration needed to half-saturate a low

concentration of MHC class I molecules). Anchor position

values were calculated by the equation ∑(RB � 1)2.

PSCPL peptides were synthesized using standard solid-

phase Fmoc chemistry on 2-chlorotrityl chloride resins.

Briefly, an equimolar mixture of 19 of the common Fmoc

AAs (excluding cysteine) was prepared for each synthesis

and used for coupling in eight positions, whereas a single

type of Fmoc AA (including cysteine) was used in one

position. This position was changed in each synthesis

starting with the N-terminus and ending with the C-

terminus. In one synthesis, the AA pool was used in all nine

positions. The library therefore consisted of

20 9 9 + 1 = 181 individual peptide libraries:

• 20 PSCPL sublibraries describing position 1: AX8, CX8,

DX8, … YX8

• 20 PSCPL sublibraries describing position 2: XAX7, XCX7,

XDX7, … XYX7

• etc.

• 20 PSCPL sublibraries describing position 9: X8A, X8C,

X8D, … X8Y

• a completely random peptide library: X9

X denotes the random incorporation of AA from the mixture,

whereas the single-letter AA abbreviation is used to denote

the identity of the fixed AA. Following synthesis, the peptides

were cleaved from the resin in trifluoroacetic acid/1,2-

ethanedithiol/triisopropylsilane/water 95:2:1:3 v/v/v/v,

precipitated in cold diethyl ether and extracted with water

before desaltingonC18-columns, freeze dryingandweighing.

ELISA for affinity of peptide binding

Peptides were tested for their ability to form complexes with

the SLA-1*0401 and SLA-2*0401 proteins using a previ-

ously described immunosorbent assay (Sylvester-Hvid et al.

2002). Complexes of heavy chain with human b2m and

peptide were formed in 96-well plates (Thermo Scientific).

All peptides were dissolved in 150 ll standard folding buffer

[PBS, 0.1% pluriol (Lutrol F-68, BASF), pH 7.0] using

sonication for 5 min. A fivefold dilution series of peptide in

12 consecutive points was performed in standard folding

buffer. Following a 48-h incubation (18 °C), samples were

transferred to streptavidine-coated plates (Nunc, DK-4000)

and incubated for 3 h at 4 °C. Plates were washed in wash

buffer (PBS, 0.05% Tween-20) and probed with a mouse

anti-hb2m mAb (Brodsky et al. 1979), BBM1 (70 nM), for

1 h at room temperature. All wells were washed in wash

buffer and detection antibody [HRP-conjugated goat-anti-

mouse IgG (1.13 nM); Sigma] was added followed by a 1-h

incubation at room temperature. Plates were washed and

substrate (TMB; Kirkegaard & Perry Laboratories) was

added to all wells. Reactions were stopped using H2SO4

(0.3 M) and plates read at 450 nm using a ELx808

Microplate reader (Biotek). Data were analyzed using EXCEL

and PRISM software.

A pre-folded, biotinylated FLPSDYFPSV/HLA-A*02:01(Kast et al. 1994) complex was used as a standard to

convert OD450 values to the amount of complex formed

using the second-order polynomial (Y = a + bX + cX2),

thereby enabling a direct conversion of the actual peptide

concentration offered to the actual concentration of cor-

rectly formed pMHC complex. Because the effective concen-

tration of MHC (2–5 nM) used in these assays is below the

equilibrium dissociation constant (KD) of most high-affinity

peptide–MHC interactions, the peptide concentration, ED50,

leading to half-saturation of the MHC is a reasonable

approximation of the affinity of the interaction.

Results

Analysis of the peptide-binding specificity ofSLA-2*0401 with the PSCPL

The porcine MHC class I heavy chain, SLA-2*0401, was

tested for binding PSCPL using human b2m to support

folding, as previously described (Stryhn et al. 1996; Peder-

sen et al. 2011). Similar to the SLA-1*0401 molecule

(Pedersen et al. 2011), no differences were seen for the SLA-

2*0401 allele when comparing human b2m and porcine

b2m in supporting MHC peptide binding and complex

folding (data not shown). Each peptide sublibrary was

titrated to determine the EC50 (KD) for SLA-2*0401. Thepeptide-binding preferences for the SLA-2*0401 heavy

chain protein are shown in Table 1. Amino acids in each

peptide position were assigned a RB value as described in

Materials and Methods. Amino acid residues with RB values

above 2 (shown in boldface, Table 1) indicate support of

peptide binding, whereas AAs having RB values below 0.5

(shown in italics, Table 1) are considered disfavored. A


Viral peptides bound by swine class I 3

primary AP is characterized by AAs being strongly pre-

ferred, and many AAs were unaccepted by the class I

molecule in question. Anchor position values of 15 or more

indicate primary anchors (underlined values, Table 1)

(Lamberth et al. 2008). Such anchors are of critical impor-

tance for peptide binding.

The PSCPL analysis of SLA-2*0401 identifies positions

2 and 9 as APs, as is commonly seen for many class I

molecules (Sette & Sidney 1999; Lamberth et al. 2008). The

motif in Table 1 shows that the AAs asparagine (N) and

tryptophan (W) are favored for peptide binding in positions 2

and 9 respectively. The preference of AAs with larger side

chains such as tryptophan (W), tyrosine (Y) and phenylal-

anine (F) in position 9 was revealed by RB values of 11.2,

4.1 and 3.0 respectively (Table 1). Such RB values indicate

that this anchor must align with a large and deep F pocket.

The B pocket responsible for binding the AA in position 2 of

the peptide is most stable binding hydrophilic side chains,

resulting in a preference for the AAs serine (S) and

asparagine (N) with RB values of 3.3 and 8.9 respectively.

These preferences are illustrated in Fig. 1, where logos

displaying single-letter AA codes were generated to com-

pare the in vitro SLA-binding motifs for the PSCPL with

those of in silico NetMHCpan-based peptide preference

predictions. NetMHCpan is a mathematical algorithm

developed for MHC peptide-binding predictions in several

different species of interest (Nielsen et al. 2007; Hoof et al.

2009). For SLA-1*0401, the data show that the PSCPL-

based matrix detected preferences only for aspartic acid (D)

and glutamic acid (E) in position 3 of the peptide.

Furthermore, the NetMHCpan-based logo had tyrosine (Y)

preferred in position 9 with phenylalanine (F) second most

preferred, whereas the PSCPL data indicate a preference of

phenylalanine (F) over tyrosine (Y) (Fig. 1). The PSCPL

showed the same preference for the SLA-2*0401 molecule

regarding tyrosine (Y) and phenylalanine (F) in position 9

(Fig. 1). Additionally, a shared preference for serine (S) was

seen between the two SLA-2*0401 logos in position 2, as

well as a unique preference for asparagine (N) within the

PSCPL matrix only. These differences between the NetMHC-

pan-based logo and the PSCPL-based logo likely reflect the

limited amount of peptide-binding data available for the

NetMHCpan algorithm-based predictions with regard to

MHC molecules that share specificity with these two SLA

molecules (Hoof et al. 2009). Increasing the data available

SLA-1*0401PSCPL

Bits

2

1

0

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

Bits

2

1

0

Bits

2

1

0

Bits

2

1

0

NetMHCpan NetMHCpanPSCPLSLA-2*0401

Figure 1 Sequence logo representations of the SLA-1*0401- and SLA-2*0401-binding motifs. The logos were calculated from the top 1% highest

scoring peptides selected from a pool of 200 000 natural 9mer peptides using either the positional scanning combinatorial peptide library matrix or

NetMHCpan as predictors. In the sequence logo, each peptide position is represented by a stack of letters indicating its significance for binding

(information content) and the height of each letter (amino acid) is proportional to its relative frequency (Schneider & Stephens 1990). Acidic residues

are displayed in red, basic in blue, neutral in green, and hydrophobic in black. Underrepresented amino acids are displayed upside down. SLA, swine

leukocyte antigen.

Table 1 SLA-2*0401-binding motif

SLA-2*0401

Amino acid position in peptide

1 2 3 4 5 6 7 8 9

A 0.3 0.4 1.3 1.9 0.8 0.6 2.6 0.8 0.0

C 0.4 0.1 0.6 0.9 0.2 0.4 0.5 1.0 0.0

D 0.5 0.3 0.7 0.4 1.5 2.2 1.3 1.5 0.1

E 0.7 0.1 1.1 0.8 0.8 0.1 3.7 2.7 0.0

F 2.0 0.2 0.3 0.9 1.3 1.5 0.4 0.8 3.0

G 0.1 0.9 0.4 0.2 1.0 0.2 0.4 1.6 0.1

H 1.0 0.1 0.1 0.5 0.9 2.3 0.4 0.7 0.3

I 1.2 0.7 0.6 0.5 1.2 0.8 1.7 0.6 0.3

K 0.5 0.2 0.1 1.1 0.4 0.3 0.2 0.6 0.1

L 0.5 0.3 0.9 0.8 0.8 1.0 0.7 0.7 0.1

M 2.2 0.3 1.4 0.8 1.1 1.5 1.2 1.0 0.3

N 0.8 8.9 1.6 0.7 1.8 0.3 0.8 0.7 0.1

P 0.1 0.5 0.7 3.3 1.4 1.6 0.6 0.8 0.0

Q 1.4 0.2 0.8 1.5 3.3 2.5 2.8 1.0 0.0

R 2.2 0.2 1.4 1.4 0.5 1.2 0.4 1.1 0.1

S 1.6 3.3 1.0 0.8 0.9 0.6 0.8 0.9 0.1

T 0.9 1.8 1.1 0.8 0.5 1.2 0.4 0.7 0.1

V 1.1 0.8 0.8 0.4 0.7 0.2 0.3 0.8 0.0

W 0.9 0.2 2.1 1.0 0.5 0.9 0.3 1.0 11.2

Y 1.9 0.5 3.1 1.3 0.6 0.6 0.4 1.1 4.1

Sum 20 20 20 20 20 20 20 20 20

AP value 9 77 10 9 8 10 19 5 132

Normalized relative binding (RB) values indicate whether an amino acid

is favored (RB � 2.0, bold numbers) or disfavored (RB � 0.5, italic

numbers) for binding of each position in the peptide. Anchor position

(AP) values are given by the equation ∑(RB � 1)2. Important AP values

are underlined.

SLA, swine leukocyte antigen.


Pedersen et al.4

for swine MHC class I proteins will improve the NetMHCpan

algorithm prediction profile.

Identification of peptides derived from FMDV bound bySLA-1*0401 and SLA-2*0401 alleles

The P1 region of FMDV, which codes for the viral capsid or

structural proteins, was screened for 9mer and 10mer

peptides that would potentially bind SLA-1-*0401 and

SLA-2*0401 using the NetMHCpan version 2.3 (Nielsen

et al. 2007). Nine such peptides were identified for the

SLA-1*0401 allele (four 9mers and five 10mers), and 10

peptides (five 9mers and five 10mers) were identified for

SLA-2*0401 (Table 2). Selection of peptides was based on

the NetMHCpan algorithm predictions and further analyzed

for the AA preference in positions 2, 3 and 9 of the

individual peptides based on the PSCPL-binding motifs

(Table 1 and Pedersen et al. 2011). Some peptides were

selected because of their NetMHCpan rank despite not

satisfying one of the primary anchors according to the

PSCPL motif. Others were left out despite expressing an

anchor–residue match and a proper NetMHCpan rank

because of in-house peptide limitations. Within a pool of

726 possible peptide combinations from the P1 region, eight

of the nine peptides selected for either SLA-1*0401 or SLA-

2*0401 were within the top 10 highest ranking peptides for

predicted binding (Appendix S2). All chosen peptides had a

NetMHCpan rank score of <2% (Erup et al. 2011), meaning

that <2% of random natural peptides had a predicted

binding affinity stronger than the selected peptides.

These peptides, predicted to be bound, were tested for

actual binding by SLA-1 and SLA-2 in an in vitro ELISA

binding assay. Five of the nine FMDV-derived peptides

predicted to be bound by the SLA-1*0401 molecule tested

positive in the assay. These peptides include four 9mers

(TVYNGTSKY, MTAHITVPY, SSVGVTHGY and YLS-

GIAQYY) and one 10mer (NMTAHITVPY). Additionally,

we observed that four of the 10 peptides predicted to be

bound by SLA-2*0401 produced correctly folded peptide–

MHC complexes. These included three 9mers (MTAHITVPY,

SSVGVTHGY and YLSGIAQYY) and one 10mer (NTYLS-

GIAQY). The respective binding affinity (KD) values for all

bound peptides ranged between 0.8 and 433 nM. Peptides

with KD values below 500 nM are considered strong (50–

500 nM) or very strong (0.1–50 nM) binders (Table 2).

Two additional peptides, from the VP35 region of the

Sudan Ebola virus and from the conserved hypothetical

protein of the Vibrio cholera bacteria respectively, also were

tested for binding by the SLA-2*0401 protein for compar-

ison (Table 2). These peptides previously have been shown

to be bound by SLA-1*0401 with high affinity (Pedersen

et al. 2011). As for SLA-1*0401, both peptides were found

to be strongly bound by the SLA-2*0401 molecule with KD

values of 1 and 2 nM respectively (Table 2).

Discussion

Previously, we defined the peptide motifs that are bound by

the class IMHCprotein of swine, SLA-1*0401 (Pedersen et al.

2011). We now expand this analysis to an additional SLA

protein, SLA-2*0401, and describe a method for the analysis

of a virus proteome to identify the peptides that will be bound

by porcine MHC I proteins. Peptides from the P1 region of the

FMDV genome, which encodes the structural proteins of the

virus, were identified using this method for the gene products

of two class I alleles, SLA-1*0401 and SLA-2*0401. Thesetwo class I proteins are commonly expressed alleles in

commercial pig breeds as well as in different porcine cell lines

Table 2 Peptide candidates for SLA-1*0401 and SLA-2*0401 affinity analysis

Peptide sequence

FMDV peptides tested for SLA-1*0401 FMDV peptides tested for SLA-2*0401

NetMHCpan

predicted%

rank score KD (nM)

Actual binding

status

NetMHCpan

predicted%

rank score KD (nM)

Actual binding

status

TVYNGTSKY 0.17 87 Intermediate 0.05 NB Non-binder

MTAHITVPY 0.01 0.8 Strong 0.05 7 Strong

SSVGVTHGY 0.10 433 Intermediate 0.40 14 Strong

AAKHMSNTY 1.00 ND ND 0.40 NB Non-binder

YLSGIAQYY 1.50 17 Strong 0.25 42 Strong

LAAKHMSNTY 0.20 NB Non-binder 0.12 NB Non-binder

NMTAHITVPY 2.0 67 Strong 0.80 NB Non-binder

NTYLSGIAQY 0.20 NB Non-binder 0.12 22 Strong

ATVYNGTSKY 0.15 NB Non-binder 0.80 NB Non-binder

QSSVGVTHGY 0.80 NB Non-binder 1.50 NB Non-binder

Peptide sequence

SLA-2*0401 Ebola/Cholera peptides

NetMHCpan predicted% rank score KD (nM) Actual binding status

ASYQFQLPY 0.01 2 Strong

ATAAATEAY 1.00 1 Strong

ND, not determined; NB, non-binding peptide having a KD > 5000 nM; FMDV, foot-and-mouth disease virus; SLA, swine leukocyte antigen.



used for experimental research (Smith et al. 2005; Ho et al.

2006, 2009). Hence, the determination of binding motifs for

such proteins is of great value as a tool in the analysis of

porcine immune responses and vaccine development.

A comparison of in silico predictions with in vitro analysis

of peptide binding, expressed as the formation of correctly

folded porcine MHC class I, supports previously published

data describing the capacity of the NetMHCpan predictor to

identify successful peptide/MHC binding beyond humans

(Hoof et al. 2009; Pedersen et al. 2011). All 12 different

peptides tested for the SLA-2*0401 molecule were predicted

by the NetMHCpan to be bound, and six formed complexes

with the SLA class I alpha chain and b2m in vitro (Table 2).

These data show the NetMHCpan predicted actual peptide

binding at a rate of 50%, based on the limited data available

to date. Of the nine peptides selected for SLA-1*0401-binding tests, all were predicted to be bound and five formed

complexes with the heavy chain during the in vitro analysis.

For the SLA-1*0401 molecule, 55.6% of the predictions

were confirmed. These results were obtained using a

prediction algorithm based on human MHC proteins. As

such, they illuminate the value of using the extensive

human database to apply to swine MHC peptide predictions

to identify potential immune epitopes.

Of the 10 peptides from the sequence of the FMDV

structural protein region identified by the algorithm, four of

the six that failed to be bound by SLA-2*0401 have random

binding scores. Contrarily, none of the four peptides thatwere

bound by SLA-2*0401 have random binding scores calcu-

lated fromthePSCPLmatrix values and theapproximation for

longer peptide-binding prediction described by Lundegaard

et al. (2008) (data not shown). This indicates that improve-

ment could be achieved by adding quantitative SLA peptide-

binding data derived using the PSCPL matrix for the

NetMHCpan neural network to refine the algorithm and

generatemore accurate predictions (Nielsen et al. 2007; Hoof

et al. 2009). In addition, one peptide (YLSGIAQYY) was

predicted not to be bound by SLA-2*0401 but in fact was

strongly bound. Considering several peptides predicted to be

bound failed to do so in vitro illustrates the need for more data

to enhance the NetMHCpan for species other than humans.

Differences in peptide motifs expressed by the two SLA

alleles, SLA-1*0401 and SLA-2*0401, are illustrated by

peptides bound by one allele but not by the other. The 10mer

peptide, NTYLSGIAQY, was bound by SLA-2*0401 with an

affinity (KD) value of 22 nM but was not bound by the SLA-

1*0401 alpha chain (Table 2). This supports the higher

preference for the AA tyrosine (Y) in position 3 (RB = 3.1) by

the SLA-2*0401 molecule and that the residue is disfavored

in the same position for SLA-1*0401 (RB = 0.5) peptide

binding. SLA-1*0401 is unique from most other alleles in

that this alpha chain has a peptide-binding anchor in

position 3 as well as in the more commonly observed APs 2

and 9 (Pedersen et al. 2011). Thus, the side chain of the AA

in position 3 factors into peptide binding, and so the

NTYLSGIAQY peptide does not bind despite the positions 2

and 9 anchors being satisfied (T: RB = 2.5 and Y: RB = 4.3

respectively). The opposite effect is seen for the 10mer

peptide NMTAHITVPY. This peptide is readily bound by SLA-

1*0401 with a KD value of 67 nM but is not bound by SLA-

2*0401. Methionine (M) is disfavored in position 2 in the

SLA-2*0401 protein, whereas it is favored by the SLA-

1*0401 peptide-binding motif (RB = 0.3 and 1.9 respec-

tively). Furthermore, threonine (T) in position 3 is less

favored for SLA-2*0401 than for SLA-1*0401.As with the differences in binding preferences, similarities

also are seen for the two SLA alpha chains analyzed when

analyzing these viral peptides. Some of the identified binding

peptides are shared, and this sharing can be explained by

the identical preferences for tyrosine (Y) in the important

position 9 anchor and for methionine (M) in position 1.

There is also the shared preference for serine (S) and

threonine (T) in the position 2 anchor (Table 1; Pedersen

et al. 2011). In addition, our data show that both SLA

proteins can bind 9mer peptides as well as 10mer peptides

with similar affinities. This observation may indicate that

peptide epitopes longer than nine AAs may be common for

these MHC proteins. Such peptides would lead to different

recognition patterns expressed by TCRs than the more often

reported 9mers. The reason for this lies in the conformation

of the peptide–MHC complex, which would differ in the way

the peptide bulges away from the groove. This is caused by

the bound peptide now having one or even two additional

AAs in the center available for TCR recognition, depending

on which peptide AA interacts with the respective anchors

within the MHC groove (Fig. 2).

9’mer:

1

2 9

10

10’mer:3 10

2 10

Figure 2 Illustration of commonly known 9mer peptide-binding vs.

three different proposed 10mer peptide-binding scenarios and the

effect on the exposure of the central amino acids. Black bars illustrate

important interactions between the peptide and the anchor position

residues in the B- and F-pockets of the major histocompatibility

complex (MHC) class I-binding groove. Numbers indicate the peptide

position of the primary MHC-interacting amino acids within the

peptide.


Pedersen et al.6

To date, most analysis has focused on 9mer peptides as

candidates for strong class I complex folding. Longer

peptides, specifically the 10mers, were included in this study

to expand the pool of possible peptides for analysis and,

hence, possible T-cell epitope candidates. For human class I

(Chen et al. 1994; Urban et al. 1994; Tynan et al. 2005),

binding of peptides other than 9mers has been readily shown

with resulting diverse peptide conformations within the HLA

class I-binding groove. Longer peptides not only provide a

larger pool of combinations of potential peptide sequences

and binding configurations (Fig. 2), but also contribute to

greater diversity in the center region of the peptide available

for TCR recognition as a result of the peptide bulging away

from the binding cleft. Hence, depending on the individual

binding pattern of the peptide, binding of a 10mer peptide

could include additional AAs between the anchors of the

bound peptide available for TCR interactions. Such a change

in conformation could have the potential to expand pMHC-

mediated activation of virus-specific CTLs.

The ability of MHC molecules to bind the same peptide in

different combinations, depending on the identity of the

peptide anchor residues, would appear to have distinct

advantages to the immune system. Based on the PSCPL RB

values of the respective pockets expressed by the SLA-

1*0401 and SLA-2*0401 molecules, and in combination

with the respective 10 AA peptides bound, the model

predicts that the two 10mer peptides bind by a position

2/position 10 (P2/P10) or P1/P10 anchor manner

(NMTAHITVPY) for SLA-1*0401 and a P3/P10 anchor

manner (NTYLSGIAQY) for SLA-2*0401 (Table 1; Pedersen

et al. 2011). These binding patterns are only speculations

and rely on a comparison of the AA in peptide positions 1–3

with the MHC preferences expressed for the individual AA

in P1–3, combined with the comparison of the AA in

peptide positions 9 and 10 with the preferences expressed

for the individual AA in P9. Future mapping of MHC class I-

binding motifs using an X10-based peptide library as

compared to exclusively X9 will further illuminate which

AA positions of the 10mer peptides account for binding

interactions between the peptide residues and the anchors

of the MHC class I protein groove.

In conclusion, we determined the binding motif of an

additional swine MHC class I molecule, SLA-2*0401. Fur-ther, the NetMHCpan prediction algorithm based on human

MHC proteins identified peptides from a viral proteome of

interest. Based on in silico predictions and in vitro verifica-

tions, we identified four FMDV peptides bound by SLA-

2*0401: MTAHITVPY, SSVGVTHGY, YLSGIAQYY and

NTYLSGIAQY; and five FMDV peptides bound by SLA-

1*0401: TVYNGTSKY, MTAHITVPY, SSVGVTHGY, YLS-

GIAQYY and NMTAHITVPY. Three of these peptides were

observed to be commonly shared between the two heavy

chain molecules (MTAHITVPY, SSVGVTHGY and YLS-

GIAQYY), whereas one peptide (NTYLSGIAQY) was bound

only by the SLA-2*0401 protein and two peptides (TVYNGT-

SKY and NMTAHITVPY) were bound only by the SLA-

1*0401 protein. All these peptides are candidate T-cell

epitopes by virtue of being strongly bound by two of the most

highly expressed porcine class I MHC proteins.

Analysis of antigen-specific, MHC I-restricted T cells using

MHC tetramersmadewith the peptides identified here showed

that at least one of these, MTAHITVPY bound by SLA-1

*0401, reacts with FMDV-specific T cells (Patch et al. 2011).

The reason for including Ebola and V. cholera peptides in this

study was an ambition of adding the SLA-2*0401 molecule

to the experimental setup of such tetramer staining analyses.

FMDV-specific porcine CTLs should not recognize either

Ebola- or V. cholera-derived peptides bound and presented by

the SLA-2*0401 molecule, and hence, these peptides should

make for good negative controls. Subsequent studies indicate

that other peptides represented in the data presented here are

T-cell epitopes, and confirmation of these preliminary results

is underway, now also including the SLA-2*0401 molecule.

Minimally, these data, together with our previous reports

(Patch et al. 2011; Pedersen et al. 2011), show that the

approach of using the NetMHCpan prediction algorithm,

developed for human MHC molecules, can be effective in

predicting MHC peptide interactions in other species.

Acknowledgements

This work was funded by CRIS #1940-32000-053-00D

from the Agricultural Research Service, USDA (WTG); by

the Danish Council for Independent Research, Technology

and Production Sciences (274-09-0281) (GJ); and by the

United States National Institutes of Health (NIH)

(HHSN266200400025C) (SB).

Conflict of interest

The authors claim no conflict of interest in the publication

of this information.

References

Bevan M.J. (1995) Antigen presentation to cytotoxic T lymphocytes

in vivo. Journal of Experimental Medicine 182, 639–41.

Brodsky F.M., Bodmer W.F. & Parham P. (1979) Characterization

of a monoclonal anti-beta 2-microglobulin antibody and its use

in the genetic and biochemical analysis of major histocompat-

ibility antigens. European Journal of Immunology 9, 536–45.

Chen Y., Sidney J., Southwood S., Cox A.L., Sakaguchi K.,

Henderson R.A., Appella E., Hunt D.F., Sette A. & Engelhard V.

H. (1994) Naturally processed peptides longer than nine amino

acid residues bind to the class I MHC molecule HLA-A2.1 with

high affinity and in different conformations. Journal of Immunol-

ogy 152, 2874–81.

Doherty P.C. & Zinkernagel R.M. (1975) H-2 compatibility is

required for T-cell-mediated lysis of target cells infected with

lymphocytic choriomeningitis virus. Journal of Experimental

Medicine 141, 502–7.



Erup L.M., Kloverpris H., Stryhn A., Koofhethile C.K., Sims S.,

Ndung’u T., Goulder P., Buus S. & Nielsen M. (2011)

HLARESTRICTOR—a tool for patient-specific predictions of HLA

restriction elements and optimal epitopes within peptides. Immu-

nogenetics 63, 43–55.

Ferre H., Ruffet E., Blicher T., Sylvester-Hvid C., Nielsen L.L., Hobley

T.J., Thomas O.R. & Buus S. (2003) Purification of correctly

oxidized MHC class I heavy-chain molecules under denaturing

conditions: a novel strategy exploiting disulfide assisted protein

folding. Protein Science 12, 551–9.

Hansen N.J., Pedersen L.O., Stryhn A. & Buus S. (2001) High-

throughput polymerase chain reaction cleanup in microtiter

format. Analytical Biochemistry 296, 149–51.

Harty J.T., Tvinnereim A.R. & White D.W. (2000) CD8 + T cell

effector mechanisms in resistance to infection. Annual Review of

Immunology 18, 275–308.

Ho C.S., Rochelle E.S., Martens G.W., Schook L.B. & Smith D.M.

(2006) Characterization of swine leukocyte antigen polymor-

phism by sequence-based and PCR-SSP methods in Meishan pigs.

Immunogenetics 58, 873–82.

Ho C.S., Franzo-Romain M.H., Lee Y.J., Lee J.H. & Smith D.M.

(2009) Sequence-based characterization of swine leucocyte

antigen alleles in commercially available porcine cell lines.

International Journal of Immunogenetics 36, 231–4.

Ho C.S., Lunney J.K., Lee J.H., Franzo-Romain M.H., Martens G.W.,

Rowland R.R. & Smith D.M. (2010a) Molecular characterization

of swine leucocyte antigen class II genes in outbred pig

populations. Animal Genetics 41, 428–32.

Ho C.S., Martens G.W., Amoss M.S. Jr, Gomez-Raya L., Beattie C.W.

& Smith D.M. (2010b) Swine leukocyte antigen (SLA) diversity in

Sinclair and Hanford swine. Developmental and Comparative

Immunology 34, 250–7.

Hoof I., Peters B., Sidney J., Pedersen L.E., Sette A., Lund O., Buus S.

& Nielsen M. (2009) NetMHCpan, a method for MHC class I

binding prediction beyond humans. Immunogenetics 61, 1–13.

KastW.M., Brandt R.M., Sidney J., Drijfhout J.W., Kubo R.T., Grey H.

M., Melief C.J. & Sette A. (1994) Role of HLA-A motifs in

identification of potential CTL epitopes in human papillomavirus

type 16 E6 and E7 proteins. Journal of Immunology 152, 3904–12.

Lamberth K., Roder G., Harndahl M., Nielsen M., Lundegaard C.,

Schafer-Nielsen C., Lund O. & Buus S. (2008) The peptide-

binding specificity of HLA-A*3001 demonstrates membership of

the HLA-A3 supertype. Immunogenetics 60, 633–643.

Lundegaard C., Lund O. & Nielsen M. (2008) Accurate approxi-

mation method for prediction of class I MHC affinities for peptides

of length 8, 10 and 11 using prediction tools trained on 9mers.

Bioinformatics 24, 1397–8.

Nielsen M., Lundegaard C., Blicher T. et al. (2007) NetMHCpan, a

method for quantitative predictions of peptide binding to any HLA-

A and -B locus protein of known sequence. PLoS ONE 2, e796.

Ostergaard P.L., Nissen M.H., Hansen N.J., Nielsen L.L., Lauen-

moller S.L., Blicher T., Nansen A., Sylvester-Hvid C., Thromsen

A.R. & Buus S. (2001) Efficient assembly of recombinant major

histocompatibility complex class I molecules with preformed

disulfide bonds. European Journal of Immunology 31, 2986–96.

Patch J.R., Pedersen L.E., Toka F.N., Moraes M., Grubman M.J.,

Nielsen M., Jungersen G., Buus S. & Golde W.T. (2011) Induction

of foot-and-mouth disease virus-specific cytotoxic T cell killing by

vaccination. Clinical and Vaccine Immunology 18, 280–8.

Pedersen L.E., Harndahl M., Rasmussen M., Lamberth K., Golde W.

T., Lund O., Nielsen M. & Buus S. (2011) Porcine major

histocompatibility complex (MHC) class I molecules and analysis

of their peptide-binding specificities. Immunogenetics 63, 821–34.

Schmidt E.G., Buus S., Thorn M., Stryhn A., Leisner C. & Claesson

M.H. (2009) Peptide specific expansion of CD8(+) T cells by

recombinant plate bound MHC/peptide complexes. Journal of

Immunological Methods 340, 25–32.

Schneider T.D. & Stephens R.M. (1990) Sequence logos: a newway to

displayconsensus sequences.NucleicAcidsResearch18, 6097–100.

Sette A. & Sidney J. (1999) Nine major HLA class I supertypes

account for the vast preponderance of HLA-A and -B polymor-

phism. Immunogenetics 50, 201–12.

Smith D.M., Lunney J.K., Martens G.W., Ando A., Lee J.H., Ho C.S.,

SchookL.,RenardC.&ChardonP. (2005)Nomenclature for factors

of the SLA class-I system, 2004. Tissue Antigens 65, 136–49.

Stryhn A., Pedersen L.O., Romme T., Holm C.B., Holm A. & Buus S.

(1996) Peptide binding specificity of major histocompatibility

complex class I resolved into an array of apparently independent

subspecificities: quantitation by peptide libraries and improved

prediction of binding. European Journal of Immunology 26, 1911–8.

Sylvester-Hvid C., Kristensen N., Blicher T., Ferre H., Lauemoller S.L.,

Wolf X.A., Lamberth K., Nissen M.H., Pedersen L.O. & Buus S.

(2002) Establishment of a quantitative ELISA capable of deter-

miningpeptide–MHCclass I interaction.TissueAntigens59, 251–8.

Tynan F.E., Borg N.A., Miles J.J. et al. (2005) High resolution

structures of highly bulged viral epitopes bound to major

histocompatibility complex class I. Implications for T-cell receptor

engagement and T-cell immunodominance. Journal of Biological

Chemistry 280, 23900–9.

Urban R.G., Chicz R.M., Lane W.S., Strominger J.L., Rehm A.,

Kenter M.J., UytdeHaag F.G., Ploegh H., Uchanska-Ziegler B. &

Ziegler A. (1994) A subset of HLA-B27 molecules contains

peptides much longer than nonamers. Proceedings of the National

Academy of Sciences of the United States of America 91, 1534–8.

YeomS.C., Park C.G., Lee B.C. & LeeW.J. (2010) SLA typing using the

PCR-SSPmethod and establishment of the SLA homozygote line in

pedigreed SNU miniature pigs. Animal Science Journal 81, 158–64.

Zinkernagel R.M. & Doherty P.C. (1975) H-2 compatibility

requirement for T-cell-mediated lysis of target cells infected with

lymphocytic choriomeningitis virus. Different cytotoxic T-cell

specificities are associated with structures coded for in H-2K or

H-2D. Journal of Experimental Medicine 141, 1427–36.

Supporting information

Additional supporting information may be found in the

online version of this article.

Appendix S1 The nucleic acid and amino acid sequence of

SLA-2*0401 expressed in E. coli.

Appendix S2 NetMHCpan version 2.3 predictions

As a service to our authors and readers, this journal

provides supporting information supplied by the authors.

Such materials are peer-reviewed and may be re-organized

for online delivery, but are not copy-edited or typeset.

Technical support issues arising from supporting informa-

tion (other than missing files) should be addressed to the

authors.


Pedersen et al.8

43

5. Paper III

An analysis of affinity and stability in the identification of peptides bound by Swine

Leukocyte Antigens (SLA) combining matrix- and NetMHCpan based peptide selection

Pedersen LE, Rasmussen M, Harndahl M, Nielsen M, Buus S, Jungersen G.

Manuscript ready for submission to Immunogenetics

In this study SLA:peptide complexes are analyzed with respect to both affinity and

complex stability for the three SLA molecules, SLA-1*0401, SLA-2*0401 and SLA-3*0401.

By combining PSCPL matrix- and NetMHCpan based peptide predictions we here

illustrate a highly efficient approach for selection of potential T cell epitopes. Such

combined predictions led to identification of more peptides bound with high affinity and

stability than each prediction method alone. In addition, we describe a correlation

between low affinity peptides and the production of pSLA complexes of low stability,

and show that peptides bound with high affinity may produce both unstable as well as

highly stable pSLA complexes.

Furthermore, these studies present the complete nonamer PSCPL binding matrix for the

SLA-3*0401 molecule, introducing the use of a novel approach for screening of such SLA

binding specificities based on pSLA complex stability instead of affinity.

An analysis of affinity and stability in the identification of peptides bound by Swine Leukocyte

Antigens (SLA) combining matrix‐ and NetMHCpan based peptide selection

Lasse Eggers Pedersen1, Michael Rasmussen2, Mikkel Harndahl2, Morten Nielsen3, Søren Buus2, 4, and Gregers

Jungersen1

1 National Veterinary Institute, Technical University of Denmark

2 Laboratory of Experimental Immunology, Faculty of Health Sciences, University of Copenhagen

3 Center for Biological Sequence Analysis, Technical University of Denmark

4 Correspondence to professor Søren Buus, Department of International Health, Immunology and Microbiology,

Blegdamsvej 3, 2200 Copenhagen N., Denmark, ([email protected]).

Abstract

The affinity and stability of peptides

bound by major histocompatibility complex (MHC)

class I molecules are important factors in presentation

of viral epitopes to cytotoxic T lymphocytes (CTLs) of

the adaptive immune system. In swine, such peptide

presentations by swine leukocyte antigens (SLA) are

crucial for immunity during most viral infections and

diseases.

The identification of peptide candidates, derived from

viral genomes of interest, for SLA binding and

presentation continues to be an exhaustive task seeing

hundreds and sometimes even thousands of different

peptide sequences available within a single viral

proteome.

In this study, we expand the analysis

of SLA binding preferences by determining the

complete nonamer SLA‐3*0401 PSCPL (positional

scanning combinatorial peptide library) binding matrix

and suggest a method to facilitate the task of specific

peptide selection. This was done by combining the

ability of such a matrix, to predict peptide candidates

for MHC binding, with that of the in silico prediction

algorithm NetMHCpan. Analyses of such peptide

binding predictions were performed for two additional

SLA molecules, the SLA‐1*0401 and SLA‐2*0401, along

with the identification of several virally derived

peptide epitope candidates. In addition, a comparison

of the affinity and stability of peptide‐SLA (pSLA)

complexes was performed and a significant

observation of peptides being bound by low affinity

also produce pSLA complexes of low stability was

made. Furthermore, it was found that peptides bound

with high affinity produce pSLA complexes of various

stabilities ranging from 0 – 200 hours in half‐life

periods.

In conclusion, the combination of

NetMHCpan and SLA PSCPL matrix prediction led to

identification of more peptides with high affinity and

stability than each prediction method alone. Applying

these methods thus constitute a highly cost‐effective

guide for epitope discovery studies and MHC binding

analysis, not only in the case of porcine immunology

but for any species of interest.

Key words: MHC, binding predictions, peptide affinity,

complex stability

Introduction

MHC class I molecules can be seen as

tell‐tales of the immune system displaying to the

surroundings endogenously derived peptides from

processed self‐ as well as non‐self proteins.

Recognition of such peptide MHC complexes (pMHC)

by specific CTLs can lead to the activation of cytotoxic

T cells. Cells infected by an intracellular pathogen will

display non‐self peptides on their cell surface hereby

signaling a state of alert to circulating CTLs leading to

subsequent elimination of the infected cells.

Peptide ‐ MHC‐I affinity is the primary

criteria for presentation of peptides by MHC class I

molecules. Once bound, the stability of the pMHC

influence the timeframe in which peptides are

displayed at the cell surface of antigen presenting cells

(APCs), and hence the stability of individual pMHC has

a proposed additional impact on the immunogenicity

of the peptides, with stably bound peptides being the

more immunogenic (Busch and Pamer 1998, Harndahl

et al. 2012, van der Burg et al. 1996).

Knowledge about the MHC‐I binding

specificity is thus paramount in epitope discovery to

limit the pool of peptides available for analysis. The

peptide binding specificities of the human leukocyte

antigen (HLA) molecule have been thoroughly studied

and several high performance prediction servers, such

as NetMHCpan, exist to predict peptide‐HLA binding

(Lundegaard et al. 2011, Nielsen et al. 2007, Nielsen et

al. 2008, Zhang et al. 2009). However, knowledge of

the peptide binding motifs of MHC‐I molecules from

other species such as livestock is limited, complicating

epitope identification.

We recently reported the peptide

binding motif of the porcine MHC‐I molecules SLA‐

1*0401 and SLA‐2*0401, identified by ELISA based

positional scanning combinatorial peptide libraries

(PSCPL) (Pedersen et al. 2011, Pedersen et al. 2012).

Here, we describe the PSCPL derived peptide binding

motif of the SLA‐3*0401 molecule using a newly

developed methodology based on pMHC‐I complex

dissociation (Rasmussen et al. 2012). Combining the

peptide predictions of the individual binding matrices

for SLA‐1*0401, SLA‐2*0401 and SLA‐3*0401 with in

silico peptide binding predictions using NetMHCpan

(Hoof et al. 2009, Lundegaard et al. 2011) version 2.4,

we predicted possible peptide binders from a library

of more than 9000 nonameric peptides to investigate

if combined predictions were more accurate than

using either of the prediction tools alone. Predicted

peptides were screened for binding and high affinity

binders were identified for SLA‐1*0401, SLA‐2*0401

and SLA‐3*0401. In addition, the stability of pSLA was

measured for peptides identified to be bound with

high affinity, showing that our recently published

peptide‐HLA stability assay can easily be modified to

measure pMHC stability in other species (Harndahl et

al. 2011, Rasmussen et al. 2012). Comparing affinity

and stability data identified several high‐affinity, and

highly stable, peptide binders for all three SLA

proteins, and a tendency of peptides being bound with

high affinity also produce highly stable pSLA

complexes. Despite this it was found that high affinity

peptides in general produce pSLA complexes of

various stabilities ranging from 0 – 200 hour half‐life

periods. Furthermore, these analyses revealed a clear

correlation between peptides bound with low affinity

also produce low stability pSLA complexes illustrating

that high affinity peptide binding is indeed a

prerequisite for stable interactions. Such data indicate

that the methodology presented here is capable of

identifying potential CD8+ T‐cell epitopes as well as

pointing out peptides which are not suitable for the

production of stable pSLA complexes.

All together these methods provide

state‐of‐the‐art tools for the prediction, analysis and

selection of peptides for MHC class I binding, hereby

paving the way to actual identification of strongly, and

stably, bound peptides with the potential of being viral

peptide epitopes. These experiments introduce an

expansion of the field of pMHC analysis now to include

livestock such as the pig.


Recombinant construct encoding the SLA‐

3*0401 protein

Recombinant SLA‐3*0401 protein was

produced as described elsewhere (Pedersen et al.

2011). Briefly, DNA encoding residues 1 to 276 of the

SLA‐3*0401 alpha chain (purchased from Genscript)

was cloned into a pET28a+ vector (Novagen) upstream

of a polytag consisting of a FXa cleavage site, a

biotinylation signal peptide (BSP) and a histidine

affinity tag (HAT). Correct constructs were verified by

sequencing (ABI Prism 3100Avant, Applied Biosystems)

(Hansen et al. 2001). The validated construct of

interest was transformed into E. coli BL21(DE3),

containing the pACYC184 expression plasmid (Avidity,

Denver, CO) encoding an IPTG inducible BirA gene to

express biotin‐ligase. This leads to almost complete in

vivo biotinylation (85 – 100%) of the desired product

(Leisner et al. 2008).

Expression and purification of recombinant SLA‐

3*0401, human‐ and porcine beta‐2 microglobulin

(β2m)

SLA‐3*0401 heavy chain proteins were

produced recombinantly as previously described

(Pedersen et al. 2011). E. coli BL21(DE3) cells

containing the SLA proteins were lyzed in a cell‐

disrupter (Constant Cell Disruptor Systems set at 2300

bar) and the released inclusion bodies were isolated

by centrifugation (Sorval RC6, 20 min, 17000 g). The

inclusion bodies were washed twice in PBS, 0.5% NP‐

40 (Sigma), 0.1% deoxycholic acid (DOC, Sigma) and

extracted into Urea‐Tris buffer (8M urea, 25mM Tris,

pH 8.0).

SLA‐3*0401 heavy chain proteins extracted in 8

M urea were purified as previously described for the

SLA‐1*0401 molecule (Pedersen et al. 2011). In brief,

this was done by successive immobilized metal

adsorption, hydrophobic interaction‐, and eventually

Superdex‐200 size exclusion chromatography.

Throughout purification and storage the MHC‐I heavy

chain proteins were dissolved in 8 M Urea to ensure

denaturation. The pre‐oxidized, biotinylated and fully

denatured proteins were stored at ‐20°C in Tris

buffered 8 M urea.

Recombinant human‐ and porcine β2m were

expressed and purified as described elsewhere (Ferre

et al. 2003, Ostergaard et al. 2001, Pedersen et al.

2011).


All peptides were purchased from Schafer‐N,

Denmark (www.schafer‐n.com). Briefly, they were

synthesized by standard 9‐fluorenylmethyloxycarbonyl

(Fmoc)‐chemistry, purified by reversed‐phase high‐

performance liquid chromatography (to at least >80%

purity, frequently 95‐99% purity), validated by mass

spectrometry, and quantified by weight.

Positional scanning combinatorial peptide

libraries (PSCPL) were used to determine the SLA‐

3*0401 binding motif. The experimental strategy of

PSCPL has previously been described for both HLA

(Stryhn et al. 1996) and SLA (Pedersen et al. 2011)

proteins. The individual PSCPL peptide stability

analysis was performed using 125Iodine‐labeled porcine

β2m (125I‐ β2m) in a FlashPlate® based proximity assay

(Harndahl et al. 2011) explained below.

ELISA peptide affinity analysis

Affinity analysis of peptide interactions with the SLA‐

1*0401 was performed using a slightly modified

version of a previously described immunosorbent

assay (Sylvester‐Hvid et al. 2002) most recently

described by Pedersen et al. (Pedersen et al. 2012).

Luminescent oxygen channeling assay (LOCI) for

peptide affinity analysis

Affinity analysis of peptide interactions with the SLA‐

3*0401 and SLA‐2*0401 were performed using a

conformation‐dependent anti‐SLA monoclonal

antibody based LOCI technology as previously

described for HLA (Harndahl et al. 2009). In brief,

recombinant, biotinylated SLA heavy chains were

diluted in buffer (0.1% Lutrol F 68 (BASF, Art. No.

S30101. Ludwigshafen, Germany) in PBS, pH 7.2) to a

concentration of 2 nM followed by addition of 15

times excess of porcine β2m and a dose titration of

peptide, and incubated at 18°C for 48 hours. pSLA

complexes were detected using streptavidin (SA)

donor beads and an anti‐SLA monoclonal antibody

(PT85A, VMRD, Pullman, Wash., USA) conjugated to

acceptor beads. Beads and pSLA complexes were

incubated at 18°C over night at a final bead

concentration of 5 µg/ml, then equilibrated to room

temperature 1 hour prior to detection by an EnVision

multilabel reader (Perkin Elmer, Boston, USA).

FlashPlate® scintillation proximity assay (SPA)

for SLA‐3*0401 motif mapping and pSLA stability

measures

The measurement of pSLA stability was done as

previously described for HLA (Harndahl et al. 2011)

using peptide and radiolabeled porcine β2m in a mix

with recombinant SLA heavy chain 100‐fold diluted in

(0.1% Lutrol F 68 (BASF, Art. No. S30101.

Ludwigshafen, Germany) in PBS, pH 7.2) . pSLA folding

reactions were carried out in streptavidin coated

scintillation microplates (FlashPlate® PLUS, Perkin

Elmer, Boston, USA). Steady‐state of the folding

reaction was reached by overnight incubation at 18°C

at conditions of 30 nM SLA heavy chain, 1 nM

(approximately 25,000 cpm) of 125I‐β2m and 10 µM

peptide. At steady‐state unlabeled porcine β2m was

added to a final concentration of 1 µM to prevent re‐

association of any dissociated 125I‐β2m, and the

temperature was raised to 37°C. Dissociation of pSLA

was monitored continuously on a scintillation

multiplate counter (TopCount NTX, Perkin Elmer)

modified to operate at 37°C. PSCPL data were

analyzed as described previously (Rasmussen et al.

2012) and the obtained binding matrix was visualized

by the Seq2Logo server (Thomsen and Nielsen 2012)

normalizing the values at each position to 1, using P‐

weighted Kullback‐Leibler logo types and an even

amino acid background distribution (0.05 for all amino

acids).

Results

Mapping of the SLA‐3*0401 peptide binding matrix

using PSCPL

We have recently reported the

peptide binding specificities of the swine leukocyte

antigens, SLA‐1*0401 and SLA‐2*0401 (Pedersen et al.

2011, Pedersen et al. 2012). These peptide binding

specificities were determined by nonameric PSCPL

analysis using ELISA. Since then, we have introduced a

new method to determine human MHC‐I binding

specificities (Rasmussen et al. 2012), based on a

pMHC‐I dissociation assay (Harndahl et al. 2011). This

method is based on the dissociation of 125I‐

radiolabeled human β2m from in situ formed pHLA

complexes. By 125I‐radiolabeling porcine β2m, we

adopted this methodology to allow PSCPL‐based

determination of the SLA‐3*0401 binding matrix,

illustrated in figure 1 and Table 1.

PSCPL analysis of the SLA‐3*0401

molecule revealed a binding motif with a primary

anchor in the position 2 (P2) of the peptide displaying

preferences for asparagine (N) and arginine (R) with

relative binding (RB) values of 5.6 and 2.3, respectively

(Figure 1 and Table 1). This suggests a polar, acidic B‐

pocket within the SLA‐3*0401 peptide binding groove.

Auxiliary anchors were observed in P1, P3, P7 and P9

with anchor position (AP) values of 9, 11, 11 and 9,

respectively. For P9 it was found that methionine (M),

phenylalanine (F) and tyrosine (Y) were preferred with

RB values of 2.4, 2.3 and 2.0, respectively (Table 1).

Isoleucine (I) and aspartic acid (D) were also accepted

in P9 having RB values of 1.8 and 1.5, respectively.

Furthermore, it was observed that arginine (R),

glutamine (Q), lysine (K) and proline (P) were all

disfavored at P9 (RB values of 0.2, 0.3, 0.3 and 0.4,

respectively). These results indicate the presence of a

hydrophobic F‐pocket within the binding groove of

SLA‐3*0401, although this pocket does not function as

a primary anchor for peptide binding which again

makes the binding motif for SLA‐3*0401 rather

atypical when compared to the dominating P2 and P9

anchors that are commonly observed for both SLA and

HLA molecules (Kubo et al. 1994, Lamberth et al. 2008,

Pedersen et al. 2011, Pedersen et al. 2012,

Rammensee et al. 1999, Rapin et al. 2008, Sette and

Sidney 1999, Yamada et al. 1999). Instead, a primary

anchor was observed only in P2 (AP value of 27), with

an auxiliary anchor expressing a preference for

tryptophan (W), leucine (L), and phenylalanine (F) in

P7 revealing a non‐polar E‐pocket. Altogether these

results indicate that for SLA‐3*0401 P2 is the most

important peptide position in regard to peptide

binding.

Analysis of peptide‐SLA affinities

SLA‐1*0401 peptide affinity studies

were performed using a previously described ELISA

(Pedersen et al. 2011), whereas affinity measurements

for SLA‐2*0401 and SLA‐3*0401 were performed by a

modified version of a previously published LOCI based

MHC‐I binding assay (Harndahl et al. 2009), using

acceptor bead conjugated antibody PT85A which

enable the capture of correctly folded pSLA on the

surface of donor beads leading to a signal read‐out

once excited by laser (Figure 2). By using a

combination of NetMHCpan (with a 2 % rank score

threshold) and PSCPL matrix based predictions (with a

threshold of 25 corresponding to 0.5 % percentile rank

score threshold) a total of 157, 72, and 76 peptides

were selected to be tested for their ability to be bound

by SLA‐1*0401, SLA‐2*0401, and SLA‐3*0401,

respectively (Table 2 and Supporting Table 1 A ‐ C).

Note. that the NetMHCpan‐2.4 prediction software

has been trained on a limited set of SLA binding data

consisting of 14 SLA‐1*0401 and one SLA‐2*0401 (all

bound) peptides. From the 157 selected peptides

tested on SLA‐1*0401, it was found that 67 peptides

were bound with high affinity having KD values of 50

nM or less, 45 peptides had intermediate binding

affinities in the range of 50 – 500 nM, 20 peptides

turned out to be bound with weak affinities of 500 –

5000 nM, and 25 peptides were identified as non‐

binding peptides with affinities >5000 nM (Table 2). In

total, it was found that, of the 157 peptides chosen for

binding studies, 112 (71.3 %) did bind with KD values

below 500 nM. It has previously been described that

peptides with KD affinity values below 500 nM are

likely to be considered as immunological relevant

(Assarsson et al. 2007). For the SLA‐2*0401 molecule,

14 peptides (20%) of the predicted 72 were found to

have KD values lower than 500 nM, whereas 20 and 32

peptides respectively were in the range of being either

weak or non‐binding peptides. Finally, 76 peptides

were tested for SLA‐3*0401 binding identifying 20

strongly bound peptides, 9 intermediate and 9 weakly

bound peptides, whereas 38 peptides were identified

not to be bound by SLA‐3*0401 (Table 2) leading to a

total of 38 % of the peptides selected for the SLA‐

3*0401 molecule being bound with affinities stronger

than 500 nM. Individual peptide sequences, origin,

prediction rank scores, and KD affinity values are

shown in Supporting Table 1 A ‐ C. Examples of LOCI

dose titration curves for affinity analysis of two

individual peptides, MTAASYARY and KRMMMNLNY,

bound by SLA‐2*0401 and SLA‐3*0401, respectively,

are shown in Supporting Figure 1 and 2.

Combining PSCPL matrix‐ and NetMHCpan based

predictions

To quantify the complementarities of

the PSCPL matrix and NetMHCpan prediction

methods, the individual predictions were split into

subsets predicted by (a) either the matrix method or

the NetMHCpan method, (b) by both the matrix and

NetMHCpan methods, or (c) exclusively by one of the

two. The result of this analysis is shown in Table 3.

From this comparison, it is seen that the two methods

are highly complementary when it comes to

identification of SLA binding peptides. Depending on

the given SLA molecule, only 45–59 % of the peptides

validated to be bound by SLA are predicted by both

methods (Table 3, Sensitivity). Also, it is apparent that

not one method is consistently more accurate in

identifying SLA peptide binders than the other. For

SLA‐2*0401 we found that only 2 of the 20 peptides

(10 %), identified exclusively by the matrix‐based

method turned out to be bound. For the exclusive

NetMHCpan predictions 9 of 25 (36%) were found to

bind. In contrast to this, the exclusive matrix‐based

predictions for the SLA‐3*0401 molecule identified 47

% binding peptides (8 of 17), whereas the NetMHCpan

exclusive predictions for this SLA molecule only

identified 11 % binding peptides (4 of 35). The

predictive positive value (true positives divided with

the predicted positive) for the two methods on the

two SLA molecules (SLA‐2*0401 and SLA‐3*0401) is

statistically significant (p<0.05 in both cases,

comparison of ratia). That is, the predictive positive

value of the NetMHCpan method is significantly higher

for the SLA‐2*0401 molecule compared to SLA‐3*0401

and vice versa for the PSCPL matrix method

emphasizing the complementarities of the two

methods, and underlining the importance of including

both to ensure optimal coverage of the binding

repertoire of SLA molecules.

pSLA stability analysis

pMHC class stability has been

proposed to be a better determinant of

immunogenicity rather than binding affinity (Busch

and Pamer 1998, Harndahl et al. 2012, Lazarski et al.

2005, Mullbacher et al. 1999, Sant et al. 2005, van der

Burg et al. 1996). In order to investigate the stability of

pSLA complexes, we modified an existing scintillation

proximity assay (SPA) based pMHC dissociation assay

(Harndahl et al. 2011) enabling us to measure pSLA

stability. Peptides binding with an affinity better than

500 nM to SLA‐1*0401, SLA‐2*0401 or SLA‐3*0401

were tested for their pSLA stability. The correlation

between pSLA affinity and stability is shown for all

three SLA molecules in Figure 3 A – C. It was found

that the pSLA stability ranged from extremely stable

(T½ of 158 hours) to fairly unstable (T½ of 0.18 hours).

The correlation between binding affinity and stability

was statistically significant for all three SLA molecules

(Spearmans rank correlations of 0.39 (SLA‐2*0401) to

0.78 (SLA‐1*0401), and with p<0.05 in all cases, exact

permutation test). In general, it was seen that

peptides bound with high affinities were also

generating highly stable pSLA complexes with T½ ≥ 1

hour (Figure 3 A – B). It was observed that SLA‐3*0401

complexes were more unstable than the equivalent

pSLA‐1*0401 and pSLA‐2*0401, having only six

peptides with stability T½ periods ranging between 1.1

and 6.7 hours (Figure 3 C + Supporting Table 1 C). In

comparison, the SLA‐1*0401 had 65 peptides with

stabilities ranging from T½ of 1 hour to more than 158

hours, and several of those had T½ of more than 10

hours (Figure 3 A + Supporting Table 1 A). For SLA‐

2*0401, 18 of the 20 peptide with measured binding

affinity stronger than 500 nM were shown to be stably

bound with T½ of more than 1 hour (Supporting Table

1 B). This difference between fraction of stable binders

between the SLA‐3*0401 and the two other SLA

molecules is highly statistical significant (p<0.01,

comparison of ratia). Note, that analysis here for SLA‐

2*0401 and SLA‐3*0401 is limited to relative few

peptides identified with binding affinities stronger

than 500 nM. Additional analyses including a larger

number of peptides would be necessary to draw a

more general conclusion in regard to those two

molecules and their pSLA complex stabilities. Such

analyses are likely to confirm the pattern observed

here leading to the identification of several more

highly stable pSLA‐2*0401 but only a few additional

stable pSLA‐3*0401 due to the unique binding

characteristics of the SLA‐3*0401 molecule expressing

only one dominating anchor in P2 (Table 1) as

discussed below. Despite this, it was still possible to

identify peptides that were bound by the SLA‐3*0401

molecule with high affinity and producing highly stable

pSLA, an example being the peptide sequence

RNMSRIFPY, derived from Encephalitozoon cuniculi

IEDB ID: 93734, (2011). This peptide is an almost

perfect match for the SLA‐3*0401 binding motif (Table

1) and was also predicted to be bound by both the

matrix and NetMHCpan prediction rank scores

(Supporting Table 1 C). Furthermore, in vitro analysis

verified this sequence as an extremely strong and

stable binding peptide with a KD value of 1 nM and a

stability half‐life of T½ = 6.7 hours (Supporting Table 1

C). Such data confirm the ability of the prediction

approach, combining specific PSCPL binding matrix

and in silico predictions, to identify potential T cell

epitopes.

Discussion

The mapping of the complete

nonamer SLA‐3*0401 peptide binding matrix revealed

a single dominating anchor in P2 of the binding groove

(Table 1). Instead of a dominating P9 anchor which is

commonly seen for other SLA and HLA proteins

(Lamberth et al. 2008, Pedersen et al. 2011, Pedersen

et al. 2012, Rammensee et al. 1999, Rapin et al. 2008,

Sette and Sidney 1999), it was found that this

molecule displayed a rather atypical auxiliary peptide

anchor for P7 within the E‐pocket of the binding

groove.

We here present state‐of‐the‐art tools for peptide

prediction, selection, and binding analysis – both in

terms of affinity and stability. These analyses were

based on combining individual SLA binding matrix

predictions with those of NetMHCpan, and further test

binding affinity and stability using homogenous, high

throughput LOCI and SPA assays, respectively. The

combination of the two individual prediction methods

resulted in an enhancement of the overall peptide

selection in the range of a 10 to 45 % increase in the

final number of peptides identified to be bound

(Supporting Table 1 A ‐ C). Confirming the advantage

gained by combining the two prediction tools is seen

for peptides TSDGFINGW and ITDYIVGYY. These

peptides were both identified to be bound with high

affinity and to form highly stable pSLA but while

TSDGFINGW only came out as a candidate ligand for

SLA‐1*0401 by the matrix‐based predictions (matrix

score of 195, NetMHCpan rank of 15.00), the

ITDYIVGYY peptide was predicted exclusively by the

NetMHCpan (matrix score of 0, NetMHCpan rank of

0.03), (Supporting Table 1 A). Hence, combining the

two prediction tools resulted in inclusion of both

peptides of which one of them would have been lost,

were the predictions only performed by one of the

predictors. Similar examples can be found in

Supporting Table 1 A – C. We observe a general trend

that the overlap between the NetMHCpan and PSCPL

matrix‐based predictions is highest for the SLA

molecule where peptide‐binding data form part of the

training data for the NetMHCpan method. This is

consistent with earlier findings (Rasmussen et al.

2012), and suggests that as data for constructing the

pan‐specific predictor are enriched with SLA peptide

binding measurements, identified as described here

using both the PSCPL and pan‐specific predictions, the

NetMHCpan method will be able to learn the SLA

binding repertoire described by both methods. By way

of example is, that the correlation between

predictions performed by the PSCPL matrix and

NetMHCpan methods, increase by 33 % when the 76

SLA‐3*0401 binding measurements obtained in this

study are included in the training data (data not

shown).

Analyzing the correlation between

pSLA affinity and stability for the three SLA proteins

revealed a general trend of peptides being bound with

high affinity, were also producing high stability pSLA

(Figure 3 A – C). Still it was found that the individual

pSLA‐3*0401 turned out to display significantly lower

T½ values than those of both the pSLA‐1*0401 and

pSLA‐2*0401. Several factors, which could cause such

a tendency, remain to be addressed in detail. Factors

such as the low number of peptides tested, limitations

in regard to the diversity of peptides in the in‐house

peptide library (comprised of some 9000 different

peptides, where no peptides have the pattern of N/R

at P2 and D at P9, as compared to “the whole

universe” of peptides) or the fact that SLA‐3*0401 has

a lack of the otherwise commonly expressed P9

anchor in regard to peptide binding, could all be

causative for the production of low stability pSLA‐

3*0401 regardless of the peptide offered. As seen in

Supporting Tables 1 A + B, a trend for peptides

satisfying both the P2 and P9 anchor of SLA‐1*0401

and SLA‐2*0401, as well as the P3 anchor of SLA‐

1*0401, is that they are bound with higher stability, as

compared to peptides that satisfy only one of the two

anchors (or only two of the three anchors in regard to

SLA‐1*0401) which tend to be bound with lower

stability. This observation is consistent with previous

observations made for the human class I molecule,

HLA‐A*02:01 (Harndahl et al. 2012). Hence, for each

anchor satisfied the stability of the bound peptide

seems to increase, which could explain why the “one‐

anchor‐only” molecule, SLA‐3*0401, produces mostly

low stable pSLA. Speculating the biological relevance,

expression of a single major anchor for P2 could be a

phenotypic preference in regard to SLA‐3*0401

leading to faster turnover rates in the sense of peptide

association and dissociation, which again would claim

pSLA‐3*0401 of lower stabilities as compared to other

molecules showing slow peptide turnovers but higher

stabilities. An SLA allele repertoire combining fast‐

unstable and slow‐stable pSLA could be an advantage

for the host in the sense of effective and diverse

protection against intruding pathogens expressing

both fast, and diverse, peptide turnovers as well as

stable peptide presentation.

Having determined the SLA‐3*0401

binding matrix and knowing how to specifically map

such matrices can be of great value when screening

for peptide candidates. Not only does it provide an

overview and insight of specific biochemical features

within the binding groove, it also contributes to the

overall predictions of possible peptide candidates for

binding. The prediction tool combination presented

here prove extremely effective in the initial steps of

identifying CTL peptide epitopes now adding swine to

the list of species that can be addressed, screening

virtually any pathogen of interest. Furthermore,

improvement of the NetMHCpan alone, which has

only been trained on a very limited number of

peptides for SLA‐1*0401 and SLA‐2*0401 (15 and 1,

respectively) and none for the SLA‐3*0401 so far,

could also benefit the process of selecting peptide

candidates for pMHC binding studies – especially

when studying MHC molecules for which peptide

binding matrixes have not yet been defined. The data

achieved in this study, as well as data recently

obtained for cattle MHC‐I (BoLA) by the group at the

University of Copenhagen, will further contribute to

the constant improvement of the NetMHCpan by data

training.

Finally, we here present several

peptides derived from bacteria and viruses such as

foot‐and‐mouth disease virus (FMDV) and Influenza A

virus which could have a great relevance for future

studies in regard to disease in livestock animals, bound

with both high affinity and stability by three of the

most commonly expressed SLA molecules worldwide

(Supporting Table 1 A ‐ C). Such peptides can be

categorized as potential CTL epitopes expressing the

ability to generate stable pSLA, in that they are likely

to be targets for circulating CTLs, eventually leading to

cell activation and elimination of pathogenic intruders.

Clarifying details concerning the peptide binding

properties of swine MHC class I molecules could

contribute tremendously to epitope discoveries as

well as to the overall validation and analysis of

currently available or newly developed livestock

vaccines. Furthermore, knowledge of SLA binding

motifs could improve current studies and surveillance

of porcine immune responses to agricultural threats

such as foot‐and‐mouth disease virus (FMDV), porcine

reproductive and respiratory syndrome virus (PRRSV),

swine influenza virus and others, in what seems to be

a continuous struggle to keep animals healthy and

protected to avoid substantial outbreaks.

Ackowledgements

The authors would like to thank Lise

Lotte Bruun Nielsen and Panchale Olsen for their

expert experimental support. This work was supported

by the Danish Council for Independent Research,

Technology and Production Sciences (274‐09‐0281).

References

http://www.immuneepitope.org. 2011. Ref Type: Internet Communication

Assarsson E., Sidney J., Oseroff C., Pasquetto V., Bui H.H., Frahm N., Brander C., Peters B., Grey H. & Sette A. (2007) A quantitative analysis of the variables affecting the repertoire of T cell specificities recognized after vaccinia virus infection. J. Immunol. 178, 7890‐7901.

Busch D.H., Pamer E.G. (1998) MHC class I/peptide stability: implications for immunodominance, in vitro proliferation, and diversity of responding CTL. J. Immunol. 160, 4441‐4448.

Ferre H., Ruffet E., Blicher T., Sylvester‐Hvid C., Nielsen L.L., Hobley T.J., Thomas O.R. & Buus S. (2003) Purification of correctly oxidized MHC class I heavy‐chain molecules under denaturing conditions: a novel strategy exploiting disulfide assisted protein folding. Protein Sci. 12, 551‐559.

Hansen N.J., Pedersen L.O., Stryhn A. & Buus S. (2001) High‐throughput polymerase chain reaction cleanup in microtiter format. Anal. Biochem. 296, 149‐151.

Harndahl M., Justesen S., Lamberth K., Roder G., Nielsen M. & Buus S. (2009) Peptide binding to HLA class I molecules: homogenous, high‐throughput screening, and affinity assays. J. Biomol. Screen. 14, 173‐180.

Harndahl M., Rasmussen M., Roder G. & Buus S. (2011) Real‐time, high‐throughput measurements of peptide‐MHC‐I dissociation using a scintillation proximity assay. J. Immunol. Methods 374, 5‐12.

Harndahl M., Rasmussen M., Roder G., Dalgaard P., I, Sorensen M., Nielsen M. & Buus S. (2012) Peptide‐MHC class I stability is a better predictor than peptide affinity of CTL immunogenicity. Eur. J. Immunol. 42, 1405‐1416.

Hoof I., Peters B., Sidney J., Pedersen L.E., Sette A., Lund O., Buus S. & Nielsen M. (2009) NetMHCpan, a method for MHC class I binding prediction beyond humans. Immunogenetics 61, 1‐13.

Kubo R.T., Sette A., Grey H.M., Appella E., Sakaguchi K., Zhu N.Z., Arnott D., Sherman N., Shabanowitz J., Michel H. & . (1994) Definition of specific peptide motifs for four major HLA‐A alleles. J. Immunol. 152, 3913‐3924.

Lamberth K., Roder G., Harndahl M., Nielsen M., Lundegaard C., Schafer‐Nielsen C., Lund O. & Buus S. (2008) The peptide‐binding specificity of HLA‐A*3001 demonstrates membership of the HLA‐A3 supertype. Immunogenetics 60, 633‐643.

Lauemoller S.L., Holm A., Hilden J., Brunak S., Holst N.M., Stryhn A., Ostergaard P.L. & Buus S. (2001) Quantitative predictions of peptide binding to MHC class I molecules using specificity matrices and anchor‐stratified calibrations. Tissue Antigens 57, 405‐414.

Lazarski C.A., Chaves F.A., Jenks S.A., Wu S., Richards K.A., Weaver J.M. & Sant A.J. (2005) The kinetic stability of MHC class II:peptide complexes is a key parameter that dictates immunodominance. Immunity. 23, 29‐40.

Leisner C., Loeth N., Lamberth K., Justesen S., Sylvester‐Hvid C., Schmidt E.G., Claesson M., Buus S. & Stryhn A. (2008) One‐pot, mix‐and‐read peptide‐MHC tetramers. PLoS. One. 3, e1678.

Lundegaard C., Lund O. & Nielsen M. (2011) Prediction of epitopes using neural network based methods. J. Immunol. Methods 374, 26‐34.

Mullbacher A., Lobigs M., Yewdell J.W., Bennink J.R., Tha H.R. & Blanden R.V. (1999) High peptide affinity for MHC class I does not correlate with immunodominance. Scand. J. Immunol. 50, 420‐426.

Nielsen M., Lundegaard C., Blicher T., Lamberth K., Harndahl M., Justesen S., Roder G., Peters B., Sette A., Lund O. & Buus S. (2007) NetMHCpan, a method for quantitative predictions of peptide binding to any HLA‐A and ‐B locus protein of known sequence. PLoS. One. 2, e796.

Nielsen M., Lundegaard C., Blicher T., Peters B., Sette A., Justesen S., Buus S. & Lund O. (2008) Quantitative predictions of peptide binding to any HLA‐DR molecule of known sequence: NetMHCIIpan. PLoS. Comput. Biol. 4, e1000107.

Ostergaard P.L., Nissen M.H., Hansen N.J., Nielsen L.L., Lauenmoller S.L., Blicher T., Nansen A., Sylvester‐Hvid C., Thromsen A.R. & Buus S. (2001) Efficient assembly of recombinant major histocompatibility complex class I molecules with preformed disulfide bonds. Eur. J. Immunol. 31, 2986‐2996.

Pedersen L.E., Harndahl M., Nielsen M., Patch J.R., Jungersen G., Buus S. & Golde W.T. (2012) Identification of peptides from foot‐and‐mouth disease virus structural proteins bound by class I swine leukocyte antigen (SLA) alleles, SLA‐1*0401 and SLA‐2*0401. Anim Genet.

Pedersen L.E., Harndahl M., Rasmussen M., Lamberth K., Golde W.T., Lund O., Nielsen M. & Buus S. (2011) Porcine major histocompatibility complex (MHC) class I molecules and analysis of their peptide‐binding specificities. Immunogenetics 63, 821‐834.

Rammensee H., Bachmann J., Emmerich N.P., Bachor O.A. & Stevanovic S. (1999) SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics 50, 213‐219.

Rapin N., Hoof I., Lund O. & Nielsen M. (2008) MHC motif viewer. Immunogenetics 60, 759‐765.

Rasmussen M., Harndahl M., Nielsen M. & Buus S. (2012) Uncovering the Peptide Binding Specificity of HLA‐C. A generic approach/strategy for determining

the specificity of MHC Class I molecules by combining experimentally determined matrices with artificial neural networks. Manuscript in preparation

Sant A.J., Chaves F.A., Jenks S.A., Richards K.A., Menges P., Weaver J.M. & Lazarski C.A. (2005) The relationship between immunodominance, DM editing, and the kinetic stability of MHC class II:peptide complexes. Immunol. Rev. 207, 261‐278.

Sette A., Sidney J. (1999) Nine major HLA class I supertypes account for the vast preponderance of HLA‐A and ‐B polymorphism. Immunogenetics 50, 201‐212.

Stryhn A., Pedersen L.O., Romme T., Holm C.B., Holm A. & Buus S. (1996) Peptide binding specificity of major histocompatibility complex class I resolved into an array of apparently independent subspecificities: quantitation by peptide libraries and improved prediction of binding. Eur. J. Immunol. 26, 1911‐1918.

Sylvester‐Hvid C., Kristensen N., Blicher T., Ferre H., Lauemoller S.L., Wolf X.A., Lamberth K., Nissen M.H., Pedersen L.O. & Buus S. (2002) Establishment of a quantitative ELISA capable of determining peptide ‐ MHC class I interaction. Tissue Antigens 59, 251‐258.

Thomsen M.C., Nielsen M. (2012) Seq2Logo: a method for construction and visualization of amino acid binding motifs and sequence profiles including sequence weighting, pseudo counts and two‐sided representation of amino acid enrichment and depletion. Nucleic Acids Res. 40, W281‐W287.

van der Burg S.H., Visseren M.J., Brandt R.M., Kast W.M. & Melief C.J. (1996) Immunogenicity of peptides bound to MHC class I molecules depends on the MHC‐peptide complex stability. J. Immunol. 156, 3308‐3314.

Yamada N., Ishikawa Y., Dumrese T., Tokunaga K., Juji T., Nagatani T., Miwa K., Rammensee H.G. & Takiguchi M. (1999) Role of anchor residues in peptide binding to three HLA‐A26 molecules. Tissue Antigens 54, 325‐332.

Zhang H., Lundegaard C. & Nielsen M. (2009) Pan‐specific MHC class I predictors: a benchmark of HLA class I pan‐specific prediction methods. Bioinformatics. 25, 83‐89.

Figures

Figure 1 Binding motif of the SLA‐3*0401 molecule determined by pMHC dissociation driven PSCPL analysis and

displayed as logos. Each logo represents a single amino acid. The larger the size of the amino acid representative

letter, the higher the affinity for binding. Positive amino acid logos indicate slow pMHC dissociation rates and high

binding affinity. Negative amino acid logos indicate fast pMHC dissociation rates and low binding affinity. The

visualization was performed by the Seq2Logo server (Thomsen and Nielsen 2012) normalizing scores for each peptide

position to one, using P‐weighted Kullback‐Leibler logo types with an even amino acid background distribution (0.05

for all amino acids).

Figure 2 Illustration of the PT85A based LOCI principle for analysis of SLA‐2*0401 and SLA‐3*0401 peptide affinity.

Streptavidin (blue) coated donor beads enable the capture of biotinylated pSLA (purple (biotin), green (MHC/β2m), red

(bound peptide)). Acceptor beads were coated with a mouse‐anti‐SLA monoclonal antibody, PT85A (carbon grey).

Recognition of the pSLA by PT85A brings the donor and acceptor beads in close proximity leading to the transfer of

singlet oxygen (1O2) from the excited donor bead to the acceptor bead, which results in detectable emission from the

acceptor bead (orange).

PT85A

pSLA

Excitation680 nm 1

2O

Singlet oxygen

Donor bead

Emission520 ‐ 620 nm

Acceptor bead

Figure 3 A ‐ C pSLA affinity versus stability for predicted peptide candidates with measured KD values below 500 nM.

A: SLA‐1*0401 (n: 110). B: SLA‐2*0401 (n: 20). C: SLA‐3*0401 (n: 29).

A. SLA-1*0401 affinity vs stability

1 10 100 1000 10000 1000000

102030405050

100

150

200

PeptidesAffinity (kD, nM)

Sta

bili

ty,

Hal

f lif

e (h

)

B. SLA-2*0401 affinity vs stability

10 100 10000

102030405050

100

150

200


Sta

bili

ty,

Hal

f lif

e (h

)

C. SLA-3*0401 affinity vs stability

0.1 1 10 100 10000

2

4

6

8


Sta

bili

ty,

Hal

f lif

e (h

)

Supporting figures

Supporting Figure 1 LOCI based titration curve for affinity analysis of the SLA‐2*0401 binding of peptide MTAASYARY

(KD 75 nM, T½ of 3.8 hours).

Supporting Figure 2 LOCI based titration curve for affinity analysis of the SLA‐3*0401 binding of peptide KRMMMNLNY

(KD 4 nM, T½ of 1.1 hours).

SLA-2

0.001 0.01 0.1 1 10 100 1000 100000.0

0.2

0.4

0.6

0.8

1.0

17874/MTAASYARY

Peptide [nM]

Fol

ded

pMH

C [

nM]

SLA-3

0.001 0.01 0.1 1 10 100 1000 100000.0

0.2

0.4

0.613719/KRMMMNLNY

Peptide [nM]

Fol

ded

pMH

C [

nM]

Tables

Table 1 The complete nonamer SLA‐3*0401 binding matrix as determined by the PSCPL approach. RB values greater

than 2 are defined as favored AAs for peptide binding (marked in bold), whereas values below 0.5 are defined as

destabilizing (marked in italics), (Lauemoller et al. 2001, Stryhn et al. 1996). As previously defined, AP‐values above 15

are considered as dominating anchors of importance illustrated by underlining (Lamberth et al. 2008).

Table 2 Peptide affinity analysis of peptides predicted for SLA binding by the NetMHCpan or PSCPL matrix prediction

methods. The total numbers of identified strong (KD ≤ 50 nM), intermediate (KD 50 ‐ 500 nM), and weak (KD 500 ‐

5000) as well as non‐binding (KD ≥ 5000 nM) peptides are indicated for each SLA molecule.

SLA-3*0401 Amino acid position in peptide

1 2 3 4 5 6 7 8 9

A 0,8 0,7 0,6 0,8 0,9 1,0 0,7 0,9 0,9

C 0,4 0,6 0,5 0,7 0,5 0,5 0,7 0,9 0,5

D 0,9 0,4 0,3 0,5 0,4 0,7 0,5 0,8 1,5

E 0,4 0,4 0,4 0,6 0,4 0,9 0,5 1,2 0,9

F 0,6 0,5 3,0 1,7 1,6 0,8 2,2 1,5 2,3

G 0,9 0,6 0,4 0,3 0,7 0,6 0,5 0,5 0,6

H 0,8 0,5 0,8 1,3 1,1 1,0 0,8 0,8 0,6

I 0,7 0,7 0,8 0,5 1,1 1,0 0,6 1,9 1,8

K 2,2 0,7 0,6 1,8 0,7 1,5 0,4 0,7 0,3

L 1,7 0,5 1,1 1,4 0,9 1,1 2,1 1,4 0,9

M 1,6 0,5 1,4 1,3 1,1 1,2 1,9 0,7 2,4

N 0,6 5,6 1,2 0,9 1,7 1,1 0,6 0,5 0,7

P 0,3 0,4 0,5 0,7 0,9 0,9 0,7 1,4 0,4

Q 0,5 0,7 0,5 1,1 0,8 1,1 0,7 0,9 0,3

R 3,1 2,3 2,8 1,8 2,4 1,5 0,7 0,7 0,2

S 0,8 1,3 0,6 0,8 0,7 1,2 0,6 0,6 0,7

T 1,1 1,3 0,7 0,8 0,6 1,1 0,6 1,0 0,5

V 1,0 1,1 1,1 0,7 0,8 0,9 0,8 1,2 1,1

W 0,6 0,5 1,2 1,0 1,0 0,8 3,4 0,9 1,2

Y 0,9 0,6 1,6 1,2 1,6 1,0 0,8 1,5 2,0

Sum 20 20 20 20 20 20 20 20 20

AP‐value 9 27 11 4 5 1 11 3 9

Molecule ≤ 50 50 ‐ 500 500 ‐ 5000 ≥ 5000

SLA‐1*0401 67 45 20 25

SLA‐2*0401 6 14 20 32

SLA‐3*0401 20 9 9 38

Peptide affinity (KD, nM)

Table 3 Comparing abilities of the PSCPL matrix‐ and NetMHCpan based prediction methods to identify true binding

peptides for SLA‐1*0401, SLA‐2*0401 and SLA‐3*0401. The sensitivity of each prediction scenario is given as (true

positives)/(actual positives), i.e. the share of all peptides actually identified to be bound (from the pool of PSCPL

matrix or NetMHCpan predicted candidates), that were also found to be so within the pool of peptides predicted by

the respective method exclusively.

PSCPL matrix or

NetMHCpan

prediction

Peptides predicted by

either PSCPL matrix or

NetMHCpan

Sensitivity

SLA‐1*0401 157 112 (71.3 %) 1.00

SLA‐2*0401 72 20 (27.8 %) 1.00

SLA‐3*0401 76 29 (38.2 %) 1.00

Combined

prediction


PSCPL matrix and

NetMHCpan

Sensitivity

SLA‐1*0401 61 53 (86.9 %) 0.473

SLA‐2*0401 27 9 (33.3 %) 0.450

SLA‐3*0401 24 17 (70.8 %) 0.586

NetMHCpan

prediction


NetMHCpan exclusivelySensitivity

SLA‐1*0401 76 44 (57.9 %) 0.393

SLA‐2*0401 25 9 (36.0 %) 0.450

SLA‐3*0401 35 4 (11.4 %) 0.138

PSCPL matrix

predictionPeptides predicted by

PSCPL matrix exclusively

Sensitivity

SLA‐1*0401 20 15 (75.0 %) 0.134

SLA‐2*0401 20 2 (10.0 %) 0.100

SLA‐3*0401 17 8 (47.1 %) 0.276

Peptides bound

Peptides bound

Peptides bound

Peptides bound

Supporting Tables

Supporting Table 1 A

SLA‐1*0401 peptide binding analysis

Prediction Affinity/Rank

Peptide amino acid sequence Origin/Uniprot ID/Position PSCPL matrix NetMHCpan Affinity (KD, nM) Stability (T1/2, h)

Predicted by PSCPL

matrix

Predicted by

NetMHCpan

AVEDFLAFF Andes virus/Q9E005/1037 679 2.00 (WB) 2 22,69 Yes Yes

STAPTGSWFNA 186 0.40 (SB) 2 7,30 Yes Yes

GTDSGFAAY SARS-like coronavirus/Q3I5J2/187 65 0.08 (SB) 2 158,75 Yes Yes

MADSFKSDY NA 170 1.00 (WB) 2 14,71 Yes YesVLDKWNTNY Variola virus/Q89167/164 81 0.80 (WB) 3 25,12 Yes YesTIDKSSPLY Vaccinia virus/Q85324/95 172 0.25 (SB) 3 16,68 Yes Yes

MIDSDEWVY Entamoeba histolytica/Q512E5/461 84 0.80 (WB) 3 49,13 Yes Yes

TSDGFINGWYersinia pestis/A6BRR8/79 195 15,00 3 5,34 Yes No

ITDITSPLW Marburg virus/P35262/2017 185 0.80 (WB) 3 65,22 Yes YesTQDLFLPFY SARS-CoV/Q202H8/55 89 3,00 3 4,21 Yes No

ITDYIVGYYEscherichia coli/B3C476/239 0 0.03 (SB) 4 13,09 No Yes

SVDGFRASY NA 353 0.40 (SB) 4 35,71 Yes Yes

KLDAWLLPFSalmonella typhi/Q8Z2M9/132 230 1.00 (WB) 4 16,26 Yes Yes

NTDAFSREYVariola virus/Q0N575/184 171 0.40 (SB) 4 6,03 Yes Yes

KLDPTNTLW Machupo virus/Q8AZ68/370 240 3,00 4 32,59 Yes No

GVEPGHAFY Mycobacterium tuberculosis/A4KFR1/150 175 0.40 (SB) 4 5,74 Yes Yes

MTAHITVPYFoot-and-mouth diseasevirus strain A24 P1protein 1 0.01 (SB) 4 7,23 No Yes

SVEVKLPDY Dengue virus type4/Q2YHF0/449 93 2.00 (WB) 5 53,56 Yes Yes

VTDPGGLYY 67 0.01 (SB) 5 12,33 Yes YesHSDDALFIY Sangassou virus/Q1L0R2/123 112 0.17 (SB) 5 8,98 Yes Yes

RADSMMLGY Pseudomonas mallei/Q62FQ6/356 131 0.10 (SB) 5 30,50 Yes Yes

EVAGAGSGF Mycobacterium tuberculosis/Q6MWX9/706 146 15,00 6 3,01 Yes No

NIDNMCHLY Vaccinia virus/Q49PZ6/18 184 1.00 (WB) 6 12,14 Yes YesFTFWTFANY NA 7 0.05 (SB) 7 38,25 No YesLIDGRTSFY Lassa virus/Q6Y630/140 213 0.30 (SB) 7 5,88 Yes Yes

YTDKIAMSYFrancisella tularensissubsp. Tularensis/A7JE81/332 5 0.03 (SB) 7 10,34 No Yes

ATDFKFAMY NA 0 0.08 (SB) 8 104,90 No Yes

RVERIKSEY Yellow fevervirus/Q9W9I1/281 85 1.50 (WB) 8 29,52 Yes Yes

SSDLRSWTF 174 2.00 (WB) 8 2,07 Yes YesTMDVNHPIY Variola virus/Q0NIE/511 342 0.80 (WB) 8 6,80 Yes YesVSDGPPTGY Variola virus/Q89222/298 144 0.50 (WB) 8 87,80 Yes YesATEDPSSGY Ebola virus/P60170/205 128 0.30 (SB) 10 4,15 Yes YesVVDALRNIY Tula virus/Q9YQR5/1005 120 0.80 (WB) 10 15,11 Yes Yes

GTDSNGMLW Mycobacterium tuberculosis/Q6MWX8/1587 226 2.00 (WB) 10 12,52 Yes Yes

ASAAHLAAY Shigella sonnei/Q3YTU7/42 80 0.03 (SB) 11 2,46 Yes YesMTAASYARY NA 4 0.01 (SB) 11 0,85 No YesIVDCLTEMY NA 145 0.40 (SB) 12 13,14 Yes YesMIEPRTLQY NA 93 0.17 (SB) 12 3,98 Yes Yes

SVAMCRTPF Yellow fevervirus/Q1X881/2440 174 3,00 12 4,14 Yes No

IVDINVKDY Influenza A virus/genbankAY702454/107 75 3,00 12 3,89 Yes No

QTDNDIWFWEscherichia coli/B3BVD0/932 71 15,00 13 36,67 Yes No

ITDITKYLY Bacillus anthracis/B0ARG1/50 96 0.05 (SB) 14 4,78 Yes Yes

YVYFYDLSY Variola virus/Q0NLZ1/504 0 0.01 (SB) 14 0,21 No YesATAAATEAY Ebola virus/Q05127/154 8 0.01 (SB) 15 6,97 No YesKSNRIPFLY Giardia lamblia/A8BAY0/318 5 0.80 (WB) 15 0,15 No YesLTDDMIAAY SARS-CoV/Q202H8/847 156 0.05 (SB) 16 11,41 Yes Yes

LSDDAVVCY SARS-like coronavirus/P0C6W6/5125 98 0.50 (WB) 16 10,38 Yes Yes

RIARFHRPY Brucella suis/Q8FVB3/115 92 0.80 (WB) 16 1,68 Yes Yes

RVSTGLYRY Mycobacterium tuberculosis/A4KJM5/1479 68 0.30 (SB) 17 3,86 Yes Yes

EIAQHGAWY Escherichia coli/B3BCL5/74 160 2.00 (WB) 18 1,14 Yes Yes

LTDSDSPTY Mycobacterium tuberculosis/A4KDY3/1298 87 0.80 (WB) 18 4,06 Yes Yes

QTDNQLAVF West Nile virus strainPTRoxo/A8KCE3/2252 135 1.50 (WB) 18 27,51 Yes Yes

QTDPLWQKYVibrio vulnificus/Q8D5T5/90 76 0.40 (SB) 20 ND Yes Yes

LLDGLLAWY Mycobacterium tuberculosis/A4KM19/247 252 1.00 (WB) 22 4,38 Yes Yes




Predicted by PSCPL

matrix

Predicted by

NetMHCpan

VVYRGTTTY SARS-like coronavirus/P0C6W6/5508 4 0.08 (SB) 28 1,25 No Yes

FTDNNELEF Bacillus anthracisTsiankovskii-I/B3J7Q9/95 149 0.50 (WB) 31 0,49 Yes Yes

CSDETTLYY Francisella tularensissubsp. Tularensis/A7JDX6/88 5 0.03 (SB) 32 1,48 No Yes

RVFPGDHFY Mycobacterium tuberculosis/A4KKI8/230 3 0.05 (SB) 32 4,07 No Yes

VSDGGPNLYInfluenza A virus(A/duck/Guangdong/173/04(H5N1))/genbank AY737302/591 190 0.08 (SB) 37 1,72 Yes Yes

SSDISFIKY Vaccinia virus/Q80HY4/66 63 0.40 (SB) 40 0,77 Yes Yes

ASYAGAGAYArtificially designedsequence 0 0.03 (SB) 40 0,36 No Yes

EVDQTKIQY Yellow fevervirus/Q9WN82/22 103 4,00 41 2,70 Yes No

CTDDNALAY SARS-like coronavirus/P0C6T7/4138 74 0.03 (SB) 43 4,59 Yes Yes

ATVKGMQSY NA 2 0.05 (SB) 50 0,36 No Yes

FTIRDVLAY Yersinia enterocolitica/A1JNP9/275 0 0.03 (SB) 52 0,36 No Yes

ASYQFQLPY Vibrio cholerae/A6Y7B0/208 4 0.01 (SB) 53 0,61 No Yes

RTDAWSYPVEscherichia coli/B3BD85/110 112 0.80 (WB) 55 1,28 Yes Yes

RSNDTELNY Marburg virus/P35262/1181 68 0.17 (SB) 57 0,24 Yes Yes

MTSGSSSGF Mycobacterium tuberculosis/Q6MWX9/2924 125 0.40 (SB) 57 0,62 Yes Yes

RLASYGLYY Variola virus/Q9PXS2/560 9 0.05 (SB) 71 0,38 No Yes

RTWAYHGSY Dengue virus type4/Q2YHF0/2788 0 0.08 (SB) 72 0,62 No Yes

SSDDIPPRW Variola virus/Q9PXS3/58 326 5,00 72 4,16 Yes No

KSDGTGTIY SARS-like coronavirus/P0C6T7/4173 296 0.40 (SB) 78 0,59 Yes Yes

RSADGSPPY Mycobacterium tuberculosis/A4KJ56/29 296 0.08 (SB) 82 0,50 Yes Yes

RTWNYHGSYWest Nile virus strainPTRoxo/A8KCE3/2831 0 0.08 (SB) 91 0,43 No Yes

QTEENLLDF NA 141 3,00 92 1,03 Yes NoALTDLGLIY Lassa virus/Q9IMI9/201 63 0.15 (SB) 93 2,08 Yes YesKSDLQPPNY NA 104 1.00 (WB) 100 0,97 Yes Yes

KMFHGGLRY Mycobacterium tuberculosis/A4KLJ6/79 1 0.08 (SB) 105 0,18 No Yes

VVAANRSAFMycobacterium tuberculosis/A4KJA4/104 78 3,00 109 0,23 Yes No

AVEGGLYPV NA 76 15,00 109 0,39 Yes NoETESVNSNY Giardia lamblia/A8BQ55/260 79 2.00 (WB) 110 0,48 Yes YesRTLASGLIY Yersinia pestis/A6BRM7/395 4 0.05 (SB) 117 0,24 No Yes

FLYPSWSLY Cryptosporidium parvum IowaII/Q5CTZ8/176 3 0.05 (SB) 122 0,34 No Yes

YAQMWTLMY Dengue virus type3/P27915/3248 0 0.08 (SB) 140 0,56 No Yes

FTYAPAGMY Listeria monocytogenes/A3FVB4/508 0 0.01 (SB) 151 0,69 No Yes

ATTFARFLYSalmonella typhi/Q8Z4W4/505 0 0.10 (SB) 155 0,26 No Yes

KTAVVVTRY Yersinia enterocolitica/A1JJ64/43 5 0.05 (SB) 165 0,72 No Yes

ASYAAAAAYArtificially designedsequence 3 0.03 (SB) 174 0,31 No Yes

ATAWRTGGY Pseudomonas mallei/Q62HR4/293 0 0.05 (SB) 174 0,51 No Yes

DTEDNVPPW NA 114 32,00 196 0,79 Yes NoYTASVVAAY NA 7 0.01 (SB) 213 0,40 No YesKVFFGPIYY Variola virus/Q0N679/1007 0 0.05 (SB) 234 1,07 No YesAVSFRNLAY NA 0 0.05 (SB) 236 0,38 No Yes

MSWESTAEY Bacillus anthracisTsiankovskii-I/B3J8I0/10 1 0.01 (SB) 261 0,54 No Yes

YTSDYFISY NA 8 0.01 (SB) 265 0,40 No YesMSAIVSCRY Salmonella typhi/Q8Z5E5/93 8 0.10 (SB) 268 0,36 No Yes

KTFEWGVFY Entamoeba histolytica/Q50Y27/4 8 0.01 (SB) 272 0,60 No Yes

YTYPCIPEY NA 8 0.01 (SB) 279 0,80 No YesFTAMQALDY Salmonella typhi/Q8Z404/38 3 0.08 (SB) 290 1,16 No Yes

RTWFHGSLY Cryptosporidium parvum IowaII/Q5CQR9/77 0 0.01 (SB) 310 0,40 No Yes

ESDMEVFDY NA 241 2.00 (WB) 312 0,36 Yes Yes

EISGSSARY Yellow fevervirus/Q9YRV3/1420 63 1.50 (WB) 324 ND Yes Yes

GSDGGLDDY NA 100 2.00 (WB) 359 0,43 Yes Yes

QTWHGDAPY Vibrio parahaemolyticus/A6B1M3/64 0 0.08 (SB) 365 0,34 No Yes

QTNLYNLLYSalmonella typhi/Q8Z632/273 2 0.03 (SB) 380 3,05 No Yes

ISAYTHWYYFrancisella tularensissubsp. Tularensis/A7JES6/146 6 0.05 (SB) 384 1,53 No Yes

AMYDPQTYY Rickettsia sibirica/Q7P8Y1/1874 2 0.05 (SB) 417 0,29 No Yes




Predicted by PSCPL

matrix

Predicted by

NetMHCpan

HSNASTLLY NA 2 0.03 (SB) 507 ND No YesKSTDSESDW NA 156 6,00 521 ND Yes NoYIFFASFYY NA 0 0.08 (SB) 541 ND No YesRTWHYCGSY NA 0 0.08 (SB) 546 ND No YesHMMAVTLFY NA 2 0.03 (SB) 734 ND No YesYTFEPHYFY NA 6 0.01 (SB) 823 ND No YesDSDDWLNKY NA 102 5,00 1001 ND Yes NoESSDDELPY NA 82 1.50 (WB) 1117 ND Yes YesVSYAAAAAY NA 2 0.03 (SB) 1200 ND No YesGTTEVNGLY NA 0 0.08 (SB) 1293 ND No YesITMVNSLTY NA 8 0.01 (SB) 1375 ND No YesWVAGVQLLY NA 4 0.08 (SB) 1950 ND No YesYANMWSLMY NA 3 0.05 (SB) 2385 ND No YesVSALRLFNY NA 92 0.40 (SB) 2429 ND Yes YesQTYMYTGQY NA 0 0.05 (SB) 3098 ND No YesITTFFTFAY NA 0 0.03 (SB) 3134 ND No YesMTRGLLGSY NA 0 0.15 (SB) 3248 ND No YesCTLNKSHLY NA 1 0.08 (SB) 3831 ND No YesVTIGNAYIY NA 5 0.10 (SB) 4233 ND No YesLTFLDCLYY NA 1 0.05 (SB) 20000 ND No YesMTRVLPFTY NA 1 0.15 (SB) 20000 ND No YesMTRGILGSY NA 0 0.12 (SB) 20000 ND No YesYSYIFLSSY NA 4 0.03 (SB) 20000 ND No YesVTRGAVLMY NA 1 0.10 (SB) 20000 ND No YesSTFTFPGIY NA 0 0.05 (SB) 20000 ND No YesATIMPHNLY NA 9 0.08 (SB) 20000 ND No YesMTRVTNNVY NA 0 0.25 (SB) 20000 ND No YesITFQSILGY NA 4 0.03 (SB) 20000 ND No YesFSIPVTFSY NA 0 0.05 (SB) 20000 ND No YesITAGYNRYY NA 2 0.10 (SB) 20000 ND No YesMTAAQIRRY NA 7 0.15 (SB) 20000 ND No YesSAYERGLRY NA 3 0.05 (SB) 20000 ND No YesGSQYVSLAY NA 1 0.08 (SB) 20000 ND No YesASFAAQLFY NA 5 0.03 (SB) 20000 ND No YesGTEYRLTLY NA 191 0.25 (SB) 20000 ND Yes YesESENISEPY NA 83 2.00 (WB) 20000 ND Yes YesTTSDFFVNY NA 105 0.08 (SB) 20000 ND Yes YesNQDLNGNWY NA 93 15,00 20000 ND Yes NoMSDIFHALV NA 98 0.80 (WB) 20000 ND Yes YesQSAANMYIY NA 69 0.17 (SB) 20000 ND Yes YesLTMDREMLY NA 66 0.01 (SB) 20000 ND Yes YesTSSARSSEW NA 189 3,00 20000 ND Yes NoNMAPEKVDF NA 76 32,00 20000 ND Yes No

LTAHYCFLY NA 1 0.01 (SB) 20000 ND No Yes

Supporting Table 1 B




Predicted by PSCPL

matrix

Predicted by

NetMHCpan

ASYQFQLPY Vibrio cholerae/A6Y7B0/208 30 0.01 (SB) 18 51,00 Yes Yes

MNYAAAAAY NA 456 0.03 (SB) 22 131,70 Yes Yes

YTIGIGAFY NA 1 0.03 (SB) 25 10,80 No Yes

SSLPSYAAYSARS-like coronavirus/P0C6T7/3924 60 0.15 (SB) 30 1,70 Yes Yes

IIYYQLAGY Variola virus/Q0N4Z8/220 174 0.40 (SB) 38 1,30 Yes Yes

ITLKVFAGY Salmonella typhi/Q83T18/304 36 0.12 (SB) 44 22,00 Yes Yes

MVASQLARYYersinia enterocolitica/A1JQV7/564 66 0.08 (SB) 56 3,90 Yes Yes

MTRGLLGSYWest Nile virus strainPTRoxo/A8KCE3/1531 1 0.03 (SB) 72 1,00 No Yes

MTAASYARY NA 55 0.01 (SB) 75 3,80 Yes Yes

QQYHRFGLY Variola virus/Q89107/120 0 0.01 (SB) 116 0,50 No Yes

YTASVVAAY NA 4 0.01 (SB) 170 1,50 No Yes

VSYAAAAAY NA 89 0.03 (SB) 186 7,80 Yes Yes

VSIPVTNTWBrucella melitensis/Q9L772/174 37 2.00 (WB) 194 5,60 Yes Yes

ESLLHQASW Ebola virus/A3RCD1/57 89 4.00 195 93,00 Yes No

FTFWTFANY NA 7 0.01 (SB) 212 2,00 No Yes

HTSALSLGYVibrio vulnificus/Q8DEM6/164 7 0.05 (SB) 219 2,90 No Yes

YTSDYFISY NA 7 0.03 (SB) 298 1,60 No Yes

MVFQNYALYListeria monocytogenes/A3FZ93/80 5 0.01 (SB) 321 6,20 No Yes

RSVWIPGRWPseudomonas mallei/Q62BF8/186 59 4.0 375 0,20 Yes No

FTIRDVLAYYersinia enterocolitica/A1JNP9/275 2 0.03 (SB) 431 1,30 No Yes

RVYNNTARY NA 101 0.08 (SB) 525 ND Yes Yes

ESPSSDEDY NA 47 32.00 718 ND Yes No

YSSPHLLRY NA 51 0.03 (SB) 729 ND Yes Yes

KSWPAAIDW NA 138 8.00 790 ND Yes No

SLRPNDIVY NA 44 1.50 (WB) 1143 ND Yes Yes

ISRQIHWCW NA 77 6.00 1464 ND Yes No




Predicted by PSCPL

matrix

Predicted by

NetMHCpan

QTYMYTGQY NA 9 0.01 (SB) 1713 ND No Yes

RYQAQQVEW NA 108 8.00 1740 ND Yes No

YANMWSLMY NA 1 0.01 (SB) 1829 ND No Yes

LTEAQDQFY NA 119 3.00 2120 ND Yes No

YTITYHDDW NA 46 5.00 2158 ND Yes No

YTNPQFNVY NA 225 0.03 (SB) 2846 ND Yes Yes

ITMVNSLTY NA 3 0.03 (SB) 3147 ND No Yes

DTRAIDQFF NA 39 5.00 3652 ND Yes No

YSRMLYIEF NA 48 1.00 (WB) 3654 ND Yes Yes

VTEPGTAQY NA 102 1.50 (WB) 4093 ND Yes Yes

MLYPRVWPY NA 1 0.03 (SB) 4427 ND No Yes

YTGPDHQEW NA 1277 8.00 5198 ND Yes No

MMAWRMMRY NA 4 0.01 (SB) 5323 ND No Yes

HTAEIQQFF NA 35 0.80 (WB) 6677 ND Yes Yes

STFATVLEY NA 1 0.01 (SB) 7274 ND No Yes

ATAVNQECW NA 56 15.00 7405 ND Yes No

SSMNSDAAY NA 83 0.40 (SB) 9871 ND Yes Yes

FAHDDRYLY NA 0 0.05 (SB) 20000 ND No Yes

FQMDYSLEY NA 1 0.03 (SB) 20000 ND No Yes

HQYPANLFY NA 1 0.01 (SB) 20000 ND No Yes

YAYNSSLLY NA 2 0.01 (SB) 20000 ND No Yes

FSSQLGLFY NA 4 0.05 (SB) 20000 ND No Yes

MTRVTNNVY NA 1 0.05 (SB) 20000 ND No Yes

HMMAVTLFY NA 1 0.01 (SB) 20000 ND No Yes

MANIFRGSY NA 2 0.05 (SB) 20000 ND No Yes

YAQMWSLMY NA 0 0.03 (SB) 20000 ND No Yes

ISVQPLWEW NA 54 6.00 20000 ND Yes No

MTAAQIRRY NA 45 0.30 (SB) 20000 ND Yes Yes

FSVPLDEGF NA 586 5.00 20000 ND Yes No

HSNLNDATY NA 132 0.40 (SB) 20000 ND Yes Yes

STEPPMLNY NA 44 1.00 (WB) 20000 ND Yes Yes

FGMPNPEGY NA 622 2.00 (WB) 20000 ND Yes Yes

YSYIFLSSY NA 39 0.01 (SB) 20000 ND Yes Yes

KSLDNYQEW NA 45 6.00 20000 ND Yes No

RVYPNPEVY NA 590 0.80 (WB) 20000 ND Yes Yes

SSVSSFERF NA 50 3.00 20000 ND Yes No

VSRLEHQMW NA 257 5.00 20000 ND Yes No

SAYYLDIGF NA 39 1.50 (WB) 20000 ND Yes Yes

QTWHGDAPY NA 48 0.25 (SB) 20000 ND Yes Yes

SSNPVMSRF NA 73 0.80 (WB) 20000 ND Yes Yes

SSNAKNSEW NA 43 5.00 20000 ND Yes No

KSAAIDGEY NA 50 1.50 (WB) 20000 ND Yes Yes

MYADDTAGW NA 44 5.00 20000 ND Yes No

Supporting Table 1 C



Peptide amino acid sequence Origin/Uniprot ID/Position PSCPL matrixNetMHCpan Affinity (KD, nM) Stability (T1/2, h)

Predicted by PSCPL

matrix

Predicted by

NetMHCpan

RNMSRIFPYEncephalitozoon cuniculi/Q8SSE7/755 287

0.10 (SB)1 6,70 Yes Yes

RVFNNYMPY SARS-CoV/Q19QX2/2140 85 0.01 (SB) 1 5,00 Yes Yes

KRMMMNLNYRift valley fevervirus/A2SZV2/1030 23

1.50 (WB)4 1,10 No Yes

RRMATTFTF Variola virus/Q0NBG5/290 31 0.80 (WB) 4 1,10 Yes Yes

RRARYWLTY Giardia lamblia/A8BPJ1/256 43 0.50 (WB) 4 1,70 Yes Yes

RAYRNALSMYellow fevervirus/Q89318/36 34

0.03 (SB)8 0,60 Yes Yes

RARKRGITMMycobacterium tuberculosis/A4KI89/219 27

0.50 (WB)11 0,30 Yes Yes

KNNFWFWEYEntamoeba histolytica/Q50ND6/107 169

0.80 (WB)12 0,90 Yes Yes

RRSRRSLTVWest Nile virus strainPTRoxo/A8KCE3/211 54

32,0017 0,40 Yes No

RIRAANLPIMycobacterium tuberculosis/A4KJM5/1391 28

3,0017 0,20 Yes No

KSFFSRLNW

Influenza A virus(A/Northern Ireland/44233/2003(H3N2))/genbank AY738729/55 21

0.30 (SB)

24 0,20 No Yes

SARRRHLVFMycobacterium tuberculosis/A4KH94/44 40

0.40 (SB)25 0,30 Yes Yes

KRIRLKHIFCryptosporidium parvum IowaII/Q5CXD6/25 40

6,0027 1,40 Yes No

YRFRFRSVY Variola virus/Q9PXS7/740 38 1.00 (WB) 33 0,40 Yes Yes

YSRPWNWTF Giardia lamblia/A8BGW5/121 24 0.30 (SB) 35 0,50 No Yes

RRFFPYYVY NA 65 0.80 (WB) 39 0,30 Yes Yes

RRRQWASCMSalmonella typhimurium/P26474/262 27

15,0040 0,40 Yes No

TSFASSWIYRickettsia rickettsii/B0BWV4/130 44

0.40 (SB)42 0,30 Yes Yes

RVRRLNWAAPseudomonas mallei/Q62BN5/105 50

32,0043 0,30 Yes No

RRLHRLLLM NA 187 1.50 (WB) 44 0,70 Yes Yes

RMFKRVFNM Yersinia pestis/A6BP81/183 43 0.03 (SB) 52 0,40 Yes Yes

LNIMNKLNI Variola virus/Q89231/129 46 32,00 89 0,20 Yes No



Peptide amino acid sequence Origin/Uniprot ID/Position PSCPL matrixNetMHCpan Affinity (KD, nM) Stability (T1/2, h)

Predicted by PSCPL

matrix

Predicted by

NetMHCpan

MARWITWAMShigella flexneri/Q7UCA1/179 29

1.50 (WB)122 0,20 Yes Yes

KTLKGGWFFVibrio vulnificus/Q8DF54/110 31

0.25 (SB)134 0,20 Yes Yes

KSYEHQTPFSARS-like coronavirus/P0C6T7/247 7

0.01 (SB)137 0,30 No Yes

KARARLLSM Salmonella typhi/Q935J1/915 300.25 (SB)

302 0,20 Yes Yes

RRFKYLLNV Machupo virus/Q6XQI0/691 75 9,00 373 0,30 Yes No

RRFNRTKPM NA 75 6,00 554 ND Yes No

SRWSRKMLM NA 37 50,00 912 ND Yes No

LNWFEIWIV NA 60 50,00 1318 ND Yes No

RIYSHIAPY NA 6 0.01 (SB) 1506 ND No Yes

RNNDPTLPY NA 75 1.50 (WB) 1659 ND Yes Yes

GMFANRWII NA 34 2.00 (WB) 2514 ND Yes Yes

RWFVRNPFF NA 22 0.17 (SB) 3034 ND No Yes

RVFKETLFL NA 27 0.80 (WB) 3758 ND Yes Yes

VTFWGFWLF NA 23 0.25 (SB) 4911 ND No Yes

RTFDRFFEE NA 31 32,00 20000 ND Yes No

RRVRRRVLV NA 66 32,00 20000 ND Yes No

RLFFIDWEY NA 55 0.25 (SB) 20000 ND Yes Yes

YTFFFTQYF NA 27 0.20 (SB) 20000 ND Yes Yes

RSFRIHILF NA 55 0.01 (SB) 20000 ND Yes Yes

RYFTVAFLF NA 23 0.20 (SB) 20000 ND No Yes

TVFYNIPPM NA 19 0.30 (SB) 20000 ND No Yes

RRRRRRAAL NA 82 7,00 20000 ND Yes No

LSNFMLWQF NA 44 4,00 20000 ND Yes No

RLRRRRHPL NA 31 3,00 20000 ND Yes No

RVFYFAIFY NA 38 0.05 (SB) 20000 ND Yes Yes

TTRHRKPTY NA 30 10,00 20000 ND Yes No

SMFDSWGPF NA 1 0.01 (SB) 20000 ND No Yes

SQYHRFPIY NA 6 0.01 (SB) 20000 ND No Yes

YMIGYTAYY NA 0 0.03 (SB) 20000 ND No Yes

RVYPNPEVY NA 7 0.03 (SB) 20000 ND No Yes

AMYDPQTYY NA 1 0.05 (SB) 20000 ND No Yes

FLYPSWSLY NA 0 0.05 (SB) 20000 ND No Yes

RQHPGLFPF NA 7 0.01 (SB) 20000 ND No Yes

VMFRNASEY NA 6 0.03 (SB) 20000 ND No Yes

HQYPANLFY NA 4 0.01 (SB) 20000 ND No Yes

YSYIFLSSY NA 1 0.03 (SB) 20000 ND No Yes

ASYAAAAAY NA 2 0.05 (SB) 20000 ND No Yes

VSYAAAAAY NA 2 0.05 (SB) 20000 ND No Yes

YAYNSSLLY NA 5 0.01 (SB) 20000 ND No Yes

RTLDTLALY NA 3 0.03 (SB) 20000 ND No Yes

FQMDYSLEY NA 3 0.03 (SB) 20000 ND No Yes

YTYATRGIY NA 3 0.05 (SB) 20000 ND No Yes

RLYPFGSYY NA 3 0.01 (SB) 20000 ND No Yes

HMMAVTLFY NA 3 0.01 (SB) 20000 ND No Yes

GMFSWNLAY NA 4 0.01 (SB) 20000 ND No Yes

SSMNSFLLY NA 5 0.01 (SB) 20000 ND No Yes

RMFLAMITY NA 10 0.01 (SB) 20000 ND No Yes

MQYLNPPPY NA 7 0.01 (SB) 20000 ND No Yes

KMFHGGLRY NA 6 0.01 (SB) 20000 ND No Yes

RLASYGLYY NA 4 0.03 (SB) 20000 ND No Yes

ITMVNSLTY NA 8 0.01 (SB) 20000 ND No Yes

MMHASTSPF NA 1 0.01 (SB) 20000 ND No Yes

Supporting Table 1 A ‐ C SLA‐1, ‐2, and ‐3*0401 peptide prediction, affinity and stability analysis as well as successful predictions. Stability analysis (T½) was

performed for peptides with a KD of 500 nM or less. Affinity (KD) is given in nano molar (nM). Stability (T½) periods are given in hours (h). NetMHCpan rank

thresholds are 0.500 for strong binders (SB) and 2.00 for weak binders (WB). Peptides with NetMHCpan rank values above 2.00 were considered as predicted

non‐binders. PSCPL matrix scores were calculated as the multiplication of all RB values for each of the nine amino acids in the peptide. A matrix score threshold

of >25 was set for peptides considered as PSCPL matrix predicted binders. Peptides with KD affinities of >500 nM are shaded (grey).

ND: Not Determined. NA: Not Assigned.

MHC molecule / total peptides tested

PSCPL Binding Matrix Affinity Measures

(KD) Stability Measures

(T½)

SLA‐1*0401 / 157 peptides Determined by ELISA (Pedersen

et al. 2011) Determined by

ELISA Determined by SPA

SLA‐2*0401 / 72 peptides Determined by ELISA (Pedersen

et al. 2012) Determined by

LOCI Determined by SPA

SLA‐3*0401 / 76 peptides Determined by SPA Determined by

LOCI Determined by SPA

Supporting Table 2 Methods used to determine the respective SLA heavy chain binding matrices as well as peptide

affinities and stabilities. Enzyme‐Linked Immunosorbent Assay (ELISA). Scintillation Proximity Assay (SPA). Luminiscent

Oxygen Channeling Assay (LOCI).

44

6. Paper IV

Induction of foot-and-mouth disease virus-specific cytotoxic T cell killing by

vaccination

Patch JR, Pedersen LE, Toka FN, Moraes M, Grubman MJ, Nielsen M, Jungersen G, Buus

S, Golde W.T.

Clin Vaccine Immunol. 2011 Feb:18(2):280-8

The work presented in this manuscript addresses the potential induction of FMDV-

specific CD8+ CTLs as a result of vaccination against FMDV. We hypothesize that using a

recombinant human adenovirus vector based vaccine, delivering FMDV antigens in a T

cell-directed manner, would enhance cytolytic activity and effectively increase pSLA

peptide concentration for stimulation of a CTL response.

The results confirm the hypothesis and show the induction of FMDV-specific CTLs

capable of killing MHC-matched target cells. These data were further confirmed by

specific SLA tetramer staining of activated PBMCs.

In this manuscript we present the first demonstration of FMDV-specific CTL killing and

confirmation by MHC tetramer staining in response to vaccination against FMDV.

CLINICAL AND VACCINE IMMUNOLOGY, Feb. 2011, p. 280–288 Vol. 18, No. 21556-6811/11/$12.00 doi:10.1128/CVI.00417-10Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Induction of Foot-and-Mouth Disease Virus-Specific CytotoxicT Cell Killing by Vaccination�§

Jared R. Patch,1 Lasse E. Pedersen,1,2,3 Felix N. Toka,1 Mauro Moraes,1 Marvin J. Grubman,1Morten Nielsen,2 Gregers Jungersen,2 Soren Buus,3 and William T. Golde1*

Plum Island Animal Disease Center, Agricultural Research Service, USDA, Greenport, New York1; National Veterinary Institute andCenter for Biological Sequence Analysis, Technical University of Denmark, Copenhagen, Denmark2; and Laboratory of

Experimental Immunology, Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark3

Received 28 September 2010/Returned for modification 5 November 2010/Accepted 10 December 2010

Foot-and-mouth disease (FMD) continues to be a significant threat to the health and economic value oflivestock species. This acute infection is caused by the highly contagious FMD virus (FMDV), which infectscloven-hoofed animals, including large and small ruminants and swine. Current vaccine strategies are alldirected toward the induction of neutralizing antibody responses. However, the role of cytotoxic T lymphocytes(CTLs) has not received a great deal of attention, in part because of the technical difficulties associated withestablishing a reliable assay of cell killing for this highly cytopathic virus. Here, we have used recombinanthuman adenovirus vectors as a means of delivering FMDV antigens in a T cell-directed vaccine in pigs. Wetested the hypothesis that impaired processing of the FMDV capsid would enhance cytolytic activity, presum-ably by targeting all proteins for degradation and effectively increasing the class I major histocompatibilitycomplex (MHC)/FMDV peptide concentration for stimulation of a CTL response. We compared such a Tcell-targeting vaccine with the parental vaccine, previously shown to effectively induce a neutralizing antibodyresponse. Our results show induction of FMDV-specific CD8� CTL killing of MHC-matched target cells in anantigen-specific manner. Further, we confirm these results by MHC tetramer staining. This work presents thefirst demonstration of FMDV-specific CTL killing and confirmation by MHC tetramer staining in response tovaccination against FMDV.

Foot-and-mouth disease (FMD) is a highly infectious viraldisease that affects cloven-hoofed animals and remains an im-portant threat to livestock throughout much of the world (re-viewed in reference 18). Many species of wild animals aresusceptible to infection, but agricultural concerns are particu-larly focused on swine, cattle, sheep, and goats. This acutedisease is characterized by fever and viremia that last 1 to 2days, the formation of vesicular lesions in the mouth and onthe feet and teats, lethargy, lameness, and loss of meat andmilk production. Clinical signs of disease resolve within 7 to10 days; however, an asymptomatic persistent infection (car-rier state) can develop, particularly in previously vaccinatedcattle (14, 17). The etiological cause of disease is FMD virus(FMDV), which has a single-stranded, positive-sense RNAgenome of 8 kb. FMDV is a member of the Picornaviridaefamily, genus Aphthovirus. Following entry of the virus into thecell and uncoating, the released viral genome serves as anmRNA for the translation of viral proteins. Initially synthe-sized as a single polypeptide, the protein is subsequentlycleaved by viral proteases, including leader (Lpro) and 3Cpro,into mature structural and nonstructural proteins (53).

FMDV is highly cytopathic, and several viral proteins areknown to play a role in disrupting the host immune response

(reviewed in reference 19). Lpro cleaves the host translationinitiation factor eIF4G, which plays a critical role in cap-de-pendent mRNA translation (12). An internal ribosomal entrysite (IRES) in the 5� untranslated region of the FMDV ge-nome allows synthesis of viral proteins to continue while ef-fectively shutting down host protein synthesis (2, 27). Lpro alsopromotes the degradation of NF-�B, a regulator of beta inter-feron (IFN-�) transcription (10). The combined actions of Lpro

allow FMDV to block synthesis of the antiviral proteins IFN-�and -�, to which FMDV is sensitive (7, 9). In addition to therole of 3Cpro in cleaving the viral polyprotein into matureprotein products, it has been shown to induce the cleavage ofeIF4A and eIF4G (3). This protease has also been implicatedin the cleavage of histone H3 (15, 54), thus further inhibitinghost transcription and translation. In addition, both the pre-cursor protein 2BC and the coexpressed mature products, 2Band 2C, are involved in disruption of the cellular secretorypathway (33, 34). Combined with the loss of host protein syn-thesis, this disruption of the secretory pathway appears to beresponsible for inhibition of new MHC class I translation andsurface expression in FMDV-infected cells (51).

The immunopathology of this acute infection includes atransient lymphopenia (1, 13). Viral infection was shown tolead to inhibition of the type I IFN response of multiple pop-ulations of dendritic cells (36, 37). In vitro, natural killer (NK)cells have been shown to lyse FMDV-infected cells (56). How-ever, analysis of cells from FMDV-infected animals shows aloss of NK cell function during the acute phase of infection(55) (reviewed in reference 16).

The development of neutralizing antibodies plays a central

* Corresponding author. Mailing address: Plum Island Animal Dis-ease Center, Agricultural Research Service, USDA, P.O. Box 848,Greenport, NY 11944. Phone: (631) 323-3249. Fax: (631) 323-3006.E-mail: [email protected].

§ Supplemental material for this article may be found at http://cvi.asm.org/.

� Published ahead of print on 22 December 2010.

280

role in FMDV immunity and often correlates with protectionof vaccinated animals against live virus challenge (32). Cur-rently available vaccines consist of killed virus in adjuvant andare effective at providing protection but do not allow for the“differentiation of infected from vaccinated animals” (DIVA)(48). A leading candidate for a next-generation FMDV vaccineis a replication-defective human adenovirus 5 (Ad5) viral vec-tor delivering FMDV capsid precursor (P1) and 3Cpro, an“empty capsid” vaccine. This platform has DIVA capabilityand is very compatible with diagnostic tests in use at present.The parent polypeptide is processed by the 3Cpro into theprotein components of the capsid (reviewed in references 20and 48). This vaccine strategy is effective in swine (30, 31, 35)and cattle (39). However, like the killed-virus vaccine andnatural infection, neutralizing antibodies raised against thevaccine strain do not provide cross-serotype protection. Thislimits the value of vaccination during an FMDV outbreak or indisease eradication (48), as virus strain matching is requiredfor vaccine efficacy.

CD8� cytotoxic T lymphocytes (CTLs) play an importantrole in adaptive immunity and are capable of directly killingvirus-infected cells in an antigen-specific manner through therelease of perforin and granzymes as well as through Fas-Fasligand interactions (reviewed in references 5 and 58). Unlikeantibodies which recognize complex antigens, CTLs recognizeantigens that have been processed and cleaved into 8- to 12-amino-acid peptides by the proteasome and then loaded intoMHC class I molecules in the lumen of the endoplasmic retic-ulum (reviewed in reference 46). These peptide/MHC com-plexes are then transported to the surface of the infected cell,where they are displayed to T cell surveillance. CTL recogni-tion occurs when an appropriate T cell receptor (TCR), inassociation with the CD8 molecule, stably binds to the peptide/MHC complex, resulting in stimulatory signaling cascadeswithin the CTL (58). The ability of CTLs to recognize epitopesthat remain hidden to neutralizing antibodies makes them at-tractive candidates for enhancing FMDV immunity.

In contrast to the focus on neutralizing antibodies, the roleof CTLs in FMDV immunity has gone largely unexplored. Theability of the virus to inhibit MHC synthesis and transport andthe speed with which the virus is cleared presumably result inlimited antigen exposure for MHC class I-restricted T cells. Inaddition to the virus-induced lymphopenia, data from in vitrostudies would predict that FMDV-specific CTLs are not elic-ited by natural FMDV infection if MHC class I expression isblocked in vivo as has been described in vitro (51). This wouldmake FMDV-infected cells invisible to class I-restricted CTLs.Given the cytopathic nature of the virus, this hypothesis hasgone largely untested, as FMDV-infected cells are rapidlylysed in vitro and cannot be used as target cells in assays forantigen-specific cell lysis.

We have overcome these limitations with two alternativeapproaches. First, we employed recombinant Ad5 FMDV vac-cine technology as a method for delivering FMDV antigens toboth stimulating and target cells for the in vitro CTL killingassay. Here, we describe our development of an assay usingswine leukocyte antigen (SLA) homozygous pigs and theMHC-matched PK (15) cells as target cells. Second, in order tomaximize the CTL response, we used a vaccine construct con-taining a mutation in 3Cpro that inhibits its ability to process P1

into mature capsids. This approach tests the hypothesis thatone result of reduced processing of the primary polypeptideprecursor would be an increase in the pool of FMDV proteinsavailable for loading of FMDV-derived peptides into MHCclass I molecules. This is then predicted to enhance stimulationof CTLs specific for FMDV peptides.

Further, to confirm that CTL killing was antigen specific andMHC restricted, we developed class I MHC tetramers usingthe NetMHCpan prediction algorithm developed for humanclass I MHC (38) and extended to other species, includingswine (reference 25 and this report). Here we report that the Tcell-targeting vaccine elicited a greater overall CTL killingresponse than the parental vaccine and correlated with theinduction of FMDV-specific CTLs by MHC tetramer staining.

MATERIALS AND METHODS

Cell lines. Porcine kidney [PK(15)] cells (ATCC CCL-33) and PK(15)-EGFPcells (described below) were maintained in minimal essential medium (MEM)(Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS)(Thermo Fisher Scientific, Waltham, MA) and 3.4 mM L-glutamine, 3 mMdextrose, 1 mM sodium pyruvate, 5% sodium bicarbonate, nonessential andessential amino acids, 2-mercaptoethanol, and antibiotics (MEM-10). Baby ham-ster kidney (BHK-21) cells (ATCC CCL-10) were maintained in basal mediumEagle (BME) (Invitrogen, Carlsbad, CA) supplemented with 10% FBS, 10%tryptose phosphate broth, 2 mM L-glutamine, and antibiotics (BME-10). Human293 cells (ATCC CRL-1573) were maintained in MEM supplemented with 10%FBS, 2 mM L-glutamine, and antibiotics. Peripheral blood mononuclear cells(PBMCs) were maintained in Roswell Park Memorial Institute (RPMI)-1640medium (Invitrogen, Carlsbad, CA) supplemented with 10% FBS, 3.4 mML-glutamine, 3 mM dextrose, 1 mM sodium pyruvate, 5% sodium bicarbonate,nonessential and essential amino acids, 2-mercaptoethanol, and antibiotics.IBRS2 (swine kidney) cells were maintained in MEM supplemented with 10%FBS, 2 mM L-glutamine, and antibiotics.

Animals. Yucatan 4b.0/4b.0 (24) pigs were obtained from Sinclair Research,Inc., Columbia, MO. Animals were held for at least 1 week before being enteredinto a study. All animal procedures were reviewed and approved by the PlumIsland Institutional Animal Care and Use Committee.

Antibodies. A mouse monoclonal antibody (MAb) against porcine CD6, cloneMIL8, was purchased from AbD Serotec (Raleigh, NC). Mouse MAbs againstporcine CD8 (R-phycoerythrin [RPE], clone 76-2-11) and CD4 (fluoresceinisothiocyanate [FITC], clone 74-12-4) were purchased from BD Biosciences (SanJose, CA). Monoclonal anti-FMDV-VP1 (clone 6HC4.1.3 [45]) and anti-VP2(VP0) (clone F1412SA [57]) were used in combination to probe for Westernblotting. In addition, a mixture of rabbit polyclonal antibodies raised against theFMDV capsid proteins (VP1, VP2, and VP3) were used for Western blotting(43). Anti-tubulin-� MAb was used for control staining of the Western blot(clone DM1A from Thermo Scientific, Fremont, CA).

Viruses. The production of viruses Ad5A24 (Ad5-FMDV-B), Ad5VSV-G(Ad5-VSV-G) (35), and Ad5-Blue (Ad5-Empty) (35) have been previously de-scribed. A mutation in the active site of the 3Cpro gene (21) was subcloned intopAd5-A24, replacing the parental, wild-type 3Cpro, and was termed pAd5A24-3C-mut. Production of this new adenovirus, Ad5A24-3C-mut (Ad5-FMDV-T),was otherwise identical to that of the parental virus. All adenoviruses were grownin 293 cells. FMDV, strain A24 Cruzeiro, was propagated in BHK-21 cells.

Western blotting. IBRS2 cells were seeded in a 96-well microplate at 20,000cells/well in 50 �l supplemented, phenol red-free MEM and incubated for 24 hat 37°C. Following attachment, cells were infected with adenovirus constructs ata multiplicity of infection (MOI) of 20, or with FMDV A24 at an MOI of 10, andthen incubated for 24 h or 5 h, respectively. Whole-cell lysates were prepared,and protein expression was evaluated by separation by SDS-PAGE followed byWestern blotting.

Vaccination. Six Yucatan 4b.0/4b.0 pigs were divided into two groups of threepigs each and vaccinated with either Ad5-FMDV-B or Ad5-FMDV-T. PBS wasadministered to one additional pig as a negative control. For both vaccinationand boosting, animals were given a total of 5 � 109 PFU of the appropriateadenovirus. Virus was delivered intradermally in a total of 2 ml to four sites onthe neck (0.5 ml/site) using the DERMA-VAC NF needle-free system (Merial,Duluth, GA). Pigs were boosted 12 weeks postvaccination and again 11 or 13months postvaccination.

VOL. 18, 2011 FMDV-SPECIFIC CYTOTOXIC T CELL KILLING 281

Isolation of PBMCs. Porcine blood was collected in Vacutainer tubes contain-ing heparin. Typically, 20 ml of blood was diluted with 14 ml of phosphate-buffered saline (PBS), followed by the addition of a 12-ml underlay of Lym-phoprep (Axis-Shield, Oslo, Norway), and then centrifuged at 1,000 � g for 20min at room temperature. The band of peripheral blood mononuclear cells(PBMCs) was harvested and washed three times with PBS, followed by suspen-sion in RPMI.

Stimulator/target cell preparation. PK(15) cells were chosen because they areMHC haplotype 4b.0/4.b0 (23), matching the MHC locus of the Yucatan swineunder study. To prepare stimulator cells, PK(15) cells were seeded into 24-wellplates and allowed to attach overnight. Cell medium was removed and cells wereinfected with adenovirus at an MOI of 10 in 0.2 ml of RPMI 1640 (nonsupple-mented) for 1 h at 37°C with occasional rocking. For target cells, we developedPK(15)-EGFP cells. pEGFP-C1 (Clontec, Mountain View, CA) was linearizedby digestion with PciI (New England BioLabs, Ipswich, MA) and transfected intoPK(15) cells using the ProFection mammalian transfection system (Promega,Madison, WI). Stably transfected cells were selected by growth in the presence ofGeneticin (500 �g/ml) (Invitrogen, Carlsbad, CA) and cloned by two rounds oflimiting dilution. For the assay, 2 � 106 cells were infected with the appropriatevirus at an MOI of 10 for 1 h in a low volume of MEM. Target cells were thengiven MEM-10 and incubated overnight at 37°C.

CD6� purification. PBMCs isolated from vaccinated pigs were suspended at aconcentration of 107 cells/ml, and 1 ml was added to each well of infected PK(15)cells (typically 10 to 16 replicate wells), followed by the addition of 1 ml/well ofRPMI 1640 (supplemented). Plates were incubated at 37°C for 3 to 5 days.Stimulated PBMCs were recovered from wells and reisolated on Lymphoprep inorder to remove dead cells. To positively select CD6� cells, purified PBMCswere labeled with mouse anti-porcine CD6 MAb (1 �l/106 cells) followed byMACS goat anti-mouse IgG microbeads and separated on LS columns using aMidiMACS magnetic separator (Miltenyi Biotec, Gladbach, Germany), accord-ing to the manufacturer’s instructions, resulting in approximately 95% CD6�

cells (data not shown).Cytotoxic T lymphocyte killing assay. In vitro expanded, CD6�-enriched cells

were suspended in supplemented RPMI 1640 at a concentration of 107 cells/ml,plated (100 �l/well) in a low-binding 96-well plate (Thermo Fisher Scientific,Waltham, MA), and diluted 2-fold to obtain final effector-to-target-cell (E:T)ratios of 50:1, 25:1, 12:1, and 6:1. Trypsinized PK(15)-EGFP target cells weresuspended at 105 cells/ml and added to the effector cells (100 �l/well). In orderto discriminate dead target cells from live cells, 5 �l of 7-amino-actinomycin D(7-AAD) (BD Biosciences, San Jose, CA) was added to each well. The plate wascentrifuged at 233 � g for 1 min and then incubated at 37°C for 4 h. Followingincubation, cells were resuspended and data were acquired by a FACSCaliburflow cytometer with a high-throughput sampler (HTS) (BD Biosciences, SanJose, CA). Data were analyzed using CellQuest Pro software (BD Biosciences,San Jose, CA) by gating on PK(15)-EGFP cells and determining the percentageof 7-AAD-positive (lysed) cells. Background lysis was determined by 7-AADstaining of target cells in the absence of effectors, and this percentage wassubtracted from all groups.

Serum neutralization assay. Endpoint titers of serum neutralization of FMDVA24 were determined according to the procedure described by Pacheco andcoworkers (40). Briefly, 4-fold serum dilutions were incubated at 37°C with 10050% tissue culture infective doses (TCID50) of FMDV serotype A24 for 1 h in amicrotiter plate, after which BHK-21 cells were added (104 cells/well). Followingincubation for 3 days, wells were scored for cytopathic effect and endpoint titerswere calculated, which represent the reciprocal of the log dilution of serum ableto neutralize 100 TCID50 of virus.

Tetramer construction and staining. Peptides and recombinant SLA-1*0401and human �2-microglobulin (�2m) were produced and complexed into tetram-ers as described previously (28) with minor modifications. Briefly, recombinant�2m and peptides were added to 10-fold and 40-fold excesses, respectively, ofSLA-1*0401 in a reaction buffer containing a cocktail of protease inhibitors [0.68mM EDTA, 4.8 �M pepstatin A, 94.4 �M 1-10 phenanthroline, 0.1 mM N-ethylmaleimide, 8.9 �M tosyl-L-lysine chloromethyl ketone (TLCK), 9.3 �Mtosyl-L-phenylalanine chloromethyl ketone (TPCK), 66.8 �M phenylmethylsul-fonyl fluoride (PMSF)] followed by the addition of streptavidin-allophycocyanin(SA-APC). For tetramer staining, 2 � 105 to 5 � 105 cells were washed with coldFACS buffer (PBS, 0.3% bovine serum albumin, 0.1% sodium azide) and resus-pended in solution containing tetramer previously mixed 1:1 with 2� FACSbuffer. Cells were incubated at room temperature with rocking for 20 min andthen washed three times in cold FACS buffer. Cells were resuspended in FACSbuffer containing fluorescent MAbs against porcine CD8 and CD4 and thenincubated at room temperature for 20 min. After being washed three times in

FACS buffer, cells were fixed in PBS containing 1% formaldehyde and stored at4°C until being analyzed by flow cytometry.

RESULTS

Mutation of FMDV 3C protease results in reduced capsidprocessing and production of anti-FMDV antibodies. Previouswork by Moraes and coworkers demonstrated that vaccinationof swine with Ad5A24 (Ad5-FMDV-B) resulted in the gener-ation of a protective anti-FMDV antibody response (35). Mu-tation of the 3C protease resulting in an amino acid change(C163S) is known to inhibit processing of P1 into mature cap-sid proteins (21). A second adenovirus construct with the sameP1 polyprotein derived from FMDV strain A24 and the mutated3C protease was made, which we designate Ad5-FMDV-T. Thisconstruct was predicted to provide enhanced stimulation of aCTL response based on previous work that demonstrated alack of antibody response to a similar construct made forFMDV strain A12 (31).

To demonstrate the reduced activity of 3C in this new con-struct, IBRS2 cells were infected with Ad5-FMDV-B, Ad5-FMDV-T, Ad5-Empty, or FMDV strain A24. Cell lysates wereanalyzed by SDS-PAGE followed by Western blotting. Wedetected the presence of P1 cleavage products VP0, VP3, andVP1 in the lysates of cells infected with either FMDV orAd5-FMDV-B, with little or no P1 precursor (Fig. 1). In con-trast, cells infected with Ad5-FMDV-T showed a significantdecrease in the presence of VP0, VP3, and VP1, with signifi-cantly more unprocessed P1 (Fig. 1), suggesting a reduction of3Cpro activity and consequent accumulation of P1.

Sera from pigs vaccinated and boosted with Ad5-FMDV-Band Ad5-FMDV-T were tested for FMDV neutralizing anti-bodies as described in Materials and Methods. All three pigsvaccinated with Ad5-FMDV-B generated neutralizing antibod-ies following priming, with elevated titers following two roundsof boosting (Fig. 2). In contrast, of the three pigs vaccinatedwith Ad5-FMDV-T, only p945 generated detectable neutraliz-ing antibodies following priming. Both p945 and p947 hadserum neutralizing antibodies following boost, with the titer forp947 remaining relatively low. Neutralizing antibodies were

FIG. 1. Reduced capsid processing of Ad5-FMDV-T. IBRS2 cellswere infected with the indicated virus as described in Materials andMethods. Following infection, cell lysates were made and proteinswere separated by SDS-PAGE and visualized by Western blottingusing either MAbs against VP2 (visualized as the VP0 precursor) andVP1 (left) or a mix of rabbit polyclonal antibodies against FMDVcapsid proteins (right).

282 PATCH ET AL. CLIN. VACCINE IMMUNOL.

never detected in serum from p925 or from the PBS negativecontrol, p863. These data suggest that Ad5-FMDV-T is lesseffective than Ad5-FMDV-B at eliciting neutralizing antibod-ies, presumably due to less efficient processing of P1 by 3Cpro.

Establishment of an FMDV CTL assay. The animals vacci-nated with either Ad5-FMDV-T, Ad5-FMDV-B, or PBS wereboosted 12 weeks postvaccination and analyzed. PBMCs werecollected 24 days postboost and stimulated for 5 days in thepresence of PK(15) cells infected with Ad5-FMDV-T and thentested for cytolytic activity against target cells that were eithermock infected or infected with Ad5-FMDV-T. Consistent withour hypothesis, we observed the greatest antigen-specific cytolysisby cells derived from pigs vaccinated with Ad5-FMDV-T (Fig. 3).Due to low effector cell numbers, we were unable to evaluatethe response against mock-infected cells for p864 and p925,and thus we omitted data from these two animals from Fig. 3.In spite of the observed elevation of antigen-specific cytolysis,with the exception of p945 we observed poor titration of targetcell killing across several effector-to-target-cell (E:T) ratios(data not shown). A similar lack of titration of cytolysis by bulkPBMCs was previously observed by Denyer and coworkers,who demonstrated that purification of CD6� cells by positiveselection prior to incubating with target cells resulted in titra-tion of cytolytic activity across E:T ratios (11).

Based on this information, we enriched for CD6� cells fol-lowing antigen stimulation in all further experiments, as de-scribed in Materials and Methods. PBMCs isolated from pigsvaccinated with Ad5-FMDV-T and enriched for CD6 expres-sion exhibited enhanced cytolysis of Ad5-FMDV-T-infectedtarget cells that titrated across E:T ratios, compared to mockor Ad5-Empty-infected cells (Fig. 4). The level of killing de-tected by this refined assay was slightly reduced, likely due tothe elimination of NK cells from the responding population.The animal vaccinated with PBS showed no specific killing (seeFig. S1 in the supplemental material). In subsequent figures,

we show the data for the 50:1 E:T ratio only, with the cytolysisof Ad5-Empty-infected cells subtracted from each group,though all four E:T ratios were always run. Finally, with thisnew assay, we tested the kinetics of the expansion culture overthe course of 5 days and found that the highest level of killing,

FIG. 2. Reduced neutralizing antibodies in Ad5-FMDV-T-vacci-nated animals. Serum samples from pigs vaccinated with PBS, Ad5-FMDV-B, or Ad5-FMDV-T were collected before and following prim-ing and boosting (solid bars, prevaccination; open bars, preboost;striped bars, pre-second boost; shaded bars, final). Serum neutralizingtiters (SNT) were determined as described in Materials and Methods.Final serum was collected 26 days or more postboost.

FIG. 3. Elevated CTL activity in Ad5-FMDV-T-vaccinated ani-mals. PBMCs were isolated from vaccinated animals, stimulated in thepresence of Ad5-FMDV-T-infected PK(15) cells for 5 days, and thenassayed for cytolytic activity against Ad5-FMDV-T-infected or mock-infected target cells, as described in Materials and Methods. The meanand standard deviation of results for duplicate wells of an E:T ratio of50:1 are shown.

FIG. 4. Titration of CD6� CTL activity. PBMCs isolated from pigsvaccinated with Ad5-FMDV-T (}, p925; f, p945; ●, p947) were stim-ulated with Ad5-FMDV-T-infected cells for 3 days and then enrichedfor CD6 surface expression. Effector and target cells were mixed at E:Tratios of 50:1, 25:1, 12:1, and 6:1. Cytolysis of target cells was measuredby flow cytometry with target cells mock infected or infected withAd5-Empty or Ad5-FMDV-T. All error bars represent the standarddeviation of results for duplicate wells.


with minimum background, was detected on day 3 (data notshown), so all expansion cultures were done for 3 days.

Specificity of CTL cytolytic activity for FMDV P1. The de-velopment of this CTL assay was achieved with these vacci-nated and boosted animals such that by 9 months following thefirst boost, antigen-specific cytolysis was difficult to detect, sug-gesting a waning of CTL immunity. Therefore, all pigs wereboosted a second time with the vaccinating virus construct(p864 to p866 with Ad5-FMDV-B and p925, p945, and p947with Ad5-FMDV-T) and activation of CTL killing was trackedusing the assay.

In order to test the ability of the different adenovirus constructsto stimulate CTLs, PBMCs isolated from pigs 4 days after boost-ing with Ad5-FMDV-B were stimulated with cells infected witheither Ad5-FMDV-B or Ad5-FMDV-T. Following enrichmentfor CD6� cells, the effector cells were mixed with target cellsinfected with either Ad5-FMDV-B, Ad5-FMDV-T, Ad5-Empty,or Ad5-VSV-G or mock infected. Ad5-VSV-G contains the genefor glycoprotein G of vesicular stomatitis virus, strain New Jersey(VSV-NJ), and served as a negative-control antigen. CD6�

cells stimulated with Ad5-FMDV-B, compared to those stim-ulated with Ad5-FMDV-T, appeared to show less overall tar-get cell killing, but there was no statistical difference betweenthese groups (Fig. 5). Specificity for FMDV P1 was shown bylittle or no cytolysis of target cells infected with Ad5-VSV-G ormock infected (Fig. 5), eliminating the possibility that expres-sion of any adenovirus-delivered transgene may contribute tocell lysis independent of antigen specificity.

Vaccine targeting T cells induces an enhanced CTL re-sponse. Following the second boost of animals, PBMCs wereobtained on days 4, 8, 14, and 28 and CTL killing activity foreach pig was assayed. Antigen-specific cytolysis remained lowthroughout the experiment in animals vaccinated with theAd5-FMDV-B construct. In Fig. 6, a small rise in antigen-specific killing on day 4 after the second boost is apparent. LowCTL lysis was observed when analyzing cells taken the day of

the boost or 8 and 14 days following the boost. These datasuggest that Ad5-FMDV-B is a poor inducer of CTLs. Datafrom all time points are shown relative to Ad5-Empty, exceptfor data for day 14, which are shown relative to Ad5-VSV-Gbecause small numbers of effector cells isolated necessitatedomitting mock and Ad5-Empty target groups.

PBMCs from pigs boosted with Ad5-FMDV-T and stimu-lated ex vivo with Ad5-FMDV-T-infected cells were tested forlysis of target cells infected with Ad5-FMDV-T, Ad5-Empty,or mock infected. In contrast to cells from the Ad5-FMDV-B-vaccinated animals, here we observed increased levels of anti-gen-specific cytolysis 4 days following the boost that waned topreboost levels by day 14 or 28 (Fig. 6). In all three animals, ahigh level of specific killing of PK(15) cells infected with theAd5-FMDV-T construct was observed. Pig 925 had little killingactivity 8 days after vaccine boost but good activity (equivalentto the day 4 result) when assayed 14 days after boost. This mayindicate a technical problem with that day’s sample from thatanimal (day 8, p925). As shown above, this animal (p925)showed no detectable neutralizing antibody either (Fig. 2).With that exception, the Ad5-FMDV-T-vaccinated animalshad higher levels of target cell killing than the Ad5-FMDV-B-vaccinated animals through 14 days following the boost.

Animals vaccinated and boosted with Ad5-FMDV-T showedelevated mean killing of Ad5-FMDV-T-infected target cellscompared to that of those vaccinated and boosted with Ad5-FMDV-B when cells isolated 4 days after the respective finalboost were analyzed. This enhanced killing was statisticallysignificant (P � 0.05; Fig. 7). As described above, the animalthat was given PBS as a negative control never showed antigen-specific killing of target cells (see Fig. S1 in the supplementalmaterial).

CTL cytolytic activity correlates with staining by an FMDVpeptide-SLA-1 tetramer. In order to further characterize theeffector cells, we developed pMHC tetramers for the SLA-1molecule expressed by these animals, SLA-1*0401. The MHCclass I peptide binding prediction algorithm NetMHCpan haspreviously been demonstrated to accurately capture the bind-ing specificity of SLA class I molecules (25). Using this pre-diction method, we identified a peptide (MTAHITVPY) de-rived from the P1 capsid precursor of FMDV strain A24 thatbinds SLA-1*0401 at a low Kd (dissociation constant) (L. E.Pedersen, M. N. Harndahl, K. Lamberth, W. T. Golde, O.Lund, M. Nielsen, and S. Buus, unpublished data). As a neg-ative control, a second SLA-1 tetramer was used that con-tained a peptide (ATAAATEAY) derived from the Ebolavirus VP35 protein, which also binds to SLA-1*0401. For theexperiments associated with the second boost (Fig. 6), a por-tion of the PBMCs isolated for evaluation of CTL killing wereset aside for SLA-1 tetramer staining.

Cells were stained with antibodies specific for CD4 and CD8and with the tetramers. Analysis by flow cytometry was carriedout by gating on the CD4/CD8� cell population and deter-mining the percentage of tetramer-positive (tetramer�) cells(see Fig. S2 in the supplemental material). For pigs vaccinatedwith Ad5-FMDV-B, little tetramer staining was detected fromsamples taken 4 days following the second boost (Fig. 8) orthereafter (data not shown). This result correlated with lowlevels of cytolytic activity from the same samples (Fig. 6). Incontrast, pigs vaccinated with Ad5-FMDV-T had a 2- to 3-fold

FIG. 5. Differential antigen stimulation. PBMCs from Ad5-FMDV-B-vaccinated pigs were isolated 4 days postboost and stimulated for 3 dayswith PK(15) cells infected with either Ad5-FMDV-T or Ad5-FMDV-B.Following CD6� purification, effector cells were assayed for cytolysis oftarget cells infected with Ad5-FMDV-T (solid bars), Ad5-FMDV-B(open bars), or Ad5-VSV-G (shaded bars) or mock-infected (striped bars)cells. Data are shown relative to Ad5-Empty. The mean and standarddeviation of results for triplicate wells are indicated.


increase, compared to preboost levels, in tetramer stainingof the CD8� population isolated 4 days following the secondboost (Fig. 8). Increased tetramer staining was correlatedwith peak cytolysis (compare Fig. 6 and Fig. 8) and was nolonger detectable by 8 days after the boost. These data showthat FMDV peptide MTAHITVPY is presented by SLA-1*0401 molecules in vivo and serve as confirmation of thepresence of anti-FMDV, MHC-restricted CTLs.

A number of other peptides from the viral capsid precursorprotein, P1, were also shown to bind the SLA-1 MHC class Iprotein. In addition, more peptides were identified that bindSLA-2 of these Yucatan pigs (SLA-2*040201) (L. E. Pedersen,M. Harndahl, M. Nielsen, J. R. Patch, G. Jungersen, S. Buus, andW. T. Golde, unpublished data). We also prepared tetramersfrom these combinations and tested them on CD8� T cells fromthe vaccinated pigs in this study. All of these tetramers werenegative for staining cells taken 4 days after boost and all othertime points tested. Only the SLA-1*0401/MTAHITVPY tet-ramer stained cells from the Ad5-FMDV-T vaccinated animals.

DISCUSSION

In the present study, we tested the ability of two nonrepli-cating adenovirus vectors to generate a CTL cytolytic response

FIG. 7. CD6� CTL activity generated by Ad5-FMDV-T vaccina-tion. Mean CTL cytolysis activity on day 4 following final boost of pigsvaccinated with Ad5-FMDV-B or Ad5-FMDV-T was determined aftersubtracting percentage of lysis of the highest negative-control targetcell group from that of Ad5-FMDV-T target cells for each individualpig. Error bars represent the standard deviations (n 3), and theasterisk indicates statistical significance (P � 0.05) as determined by anunpaired, one-tailed Student’s t test with unequal variances.

FIG. 6. CD6� CTL activity generated following second boost. PBMCs from Ad5-FMDV-B (p864, p865, and p866)- or Ad5-FMDV-T (p925,p945 and p947)-vaccinated pigs were isolated before and following boosting and treated as described in the legend to Fig. 5, except that only datafrom Ad5-FMDV-T secondary antigen stimulation are shown. Target cells: }, Ad5-FMDV-T; f, mock. Additionally, pigs 864, 865, and 866 weretested on target cells: Œ, Ad5-FMDV-B; ●, Ad5-VSV-G. Data from all time points are shown relative to Ad5-Empty. Error bars represent thestandard deviation of results for duplicate wells.


in swine to FMDV epitopes. The ability of this vaccine plat-form to induce the production of neutralizing antibodies toempty FMDV capsids in both swine (30, 31, 35, 43) and cattle(39) is well established. However, its ability to stimulate anantigen-specific CTL response has not been addressed. Wehypothesized that the parental adenovirus-based empty capsidvaccine would provide relatively poor stimulation of CTLsbecause of the efficiency of capsid processing and formation.We reasoned that inhibited processing of FMDV P1 wouldenhance CTL activation, presumably as a result of enhanceddegradation of P1 and subsequent loading of peptides ontoclass I MHC molecules. Swine were vaccinated with a T cell-directed adenovirus-based vaccine for FMDV (Ad5-FMDV-T), which contained a mutant 3Cpro that processes P1 poorly.The decreased processing of P1 by Ad5-FMDV-T was evidentin Western blot analysis, as well as in the titers of virus-neu-tralizing antibodies in serum.

Whole PBMCs isolated from vaccinated animals demon-strated cytolytic activity against appropriate target cells. How-ever, enriching for CD6� cells before mixing the effector andtarget cells improved the analysis of CTL activity. These resultsare in agreement with those of Denyer and coworkers, whodemonstrated improved titration of antigen-specific, MHC-restricted cytolysis by enriching for the CD6� population ofswine T cells (11). The CD6� cell population has been char-acterized as consisting of MHC-restricted CTLs and helper Tcells, including CD4�/CD8, CD4/CD8�, and CD4�/CD8�

populations (8, 11, 41, 42, 49). Thus, an additional advantageof positively selecting for CD6� cells is the elimination ofnon-MHC-restricted cytolytic cells such as NK cells and �� Tcells, which do not express CD6 (44).

The MHC restriction of this CTL response was stronglysupported by SLA-1 tetramer staining of cells derived fromthese animals. When peripheral blood cells derived from pigsvaccinated with Ad5-FMDV-T were analyzed, an increase intetramer�/CD4/CD8� cells was detected. The increase intetramer-positive cells correlated with the peak of cytolyticactivity, consistent with MHC-restricted cytolytic activity. Wealso noted that for pigs vaccinated with Ad5-FMDV-T, therelative tetramer staining of samples taken 4 days after thesecond boost correlates with the relative level of CTL killingactivity. However, given the small number of animals in thegroup, this correlation is not statistically significant.

The enhanced cytolysis and tetramer staining of cells derivedfrom pigs vaccinated with Ad5-FMDV-T relative to those vac-cinated with Ad5-FMDV-B are supportive of our hypothesisthat decreased capsid processing enhances CTL activity. How-ever, it is possible that the precursor and mature capsid pro-teins are cleaved differently by the proteasome, resulting in alack of MTAHITVPY peptides displayed in MHC moleculesof the Ad5-FMDV-B-infected cells. This could explain why fewtetramer-positive cells were detected in animals vaccinated withAd5-FMDV-B. At minimum, this peptide from the C-terminalhalf of the VP2 region of P1 is not near the cleavage site at the

FIG. 8. Tetramer staining of CD4/CD8� CTLs. For time points shown in Fig. 6, PBMCs were stained with antibodies against CD4 and CD8and costained with SLA-1*0401 tetramers bound to either an FMDV (red) or Ebola virus (blue) peptide. Histograms of tetramer-positive CTLsfrom Ad5-FMDV-B (A)- or Ad5-FMDV-T (B)-vaccinated animals are shown. Numbers indicate percent FMDV tetramer-positive cells from theCD8�, CD4 gate and were calculated by subtracting the percentage of cells stained with the negative-control tetramer.


junction of VP2 and VP3. Differential tetramer staining can-not, therefore, be attributed to a loss of this peptide due toprecursor cleavage by 3Cpro.

Although CTLs play a central role in the adaptive immuneresponse to intracellular antigens, their role in FMDV immu-nity has received little attention, in part because of the tech-nical difficulties in establishing appropriate assays with whichto measure their activity. In addition to the requirement forgenetic matching of MHC class I alleles between effector andtarget cells, the cytopathic nature of FMDV has hampereddevelopment of a cytolytic assay. As a consequence, investiga-tors have measured other biological properties of CD8� cells,such as proliferation (6, 26, 47, 50) or IFN-� secretion (22), asproxies for CTL killing activity. The present report describesan alternative strategy for directly assessing CTL killing inresponse to FMDV antigens and relies on recombinant expres-sion of FMDV proteins. In an earlier report, Li and coworkersdescribed vaccination of swine with an attenuated pseudora-bies virus containing a gene for the FMDV P1 capsid precursorthat was effective at generating CTL killing activity (29). How-ever, these investigators did not address MHC restriction,characterization of the cells responsible for killing, or antigenspecificity of the reported CTL activity. In contrast, we nowpresent data describing the first description of using MHCtetramers to identify antigen-specific CTLs in swine in generaland the first time that anti-FMDV CTL killing in any livestockspecies has been correlated with tetramer staining.

Whether CTLs can play a significant role in FMDV protec-tion and vaccination strategies remains to be determined. Pre-vious work in swine found that a complete lack of capsidprocessing resulted in a failure to generate neutralizing anti-bodies, and challenged animals developed clinical signs com-parable to those of unvaccinated animals (30, 31). However,the dose of vaccine used was more than a log lower than thedose used here. Reports using higher doses of a P1-basedvaccine in swine resulted in partial protection in the absence ofFMDV-specific antibodies (50). Similar results were obtainedin cattle by the same investigators (52).

In vitro, FMDV inhibition of MHC class I surface expressionbegins in as little as 30 min postinfection (51). Thus, evenunder ideal conditions where anti-FMDV CTLs are primedand expanded, control of infection by CTLs depends on hostcells infected by wild-type virus in vivo processing and display-ing FMDV peptides leading to the CTL response before inhi-bition of new MHC class I protein expression is blocked byviral proteases. Although neutralizing antibodies are likely toremain the principal focus of FMDV vaccination, CTLs may beable to contribute to cross-serotype and subtype immunity be-cause potential T cell epitopes are not limited to the structuralproteins, where most of the immune-driven diversity is located(4). Whereas antibody epitopes are large and three dimen-sional, the T cell epitopes are linear and extremely small (8 to12 amino acids) and have the potential to be highly conservedover the different serotypes of FMDV. Further work directedtoward refining the induction and measuring of the CTL re-sponse, as well as evaluating protection from challenge, isneeded to determine if induction of CTL immunity can yieldprotection across serotypes in diverse breeds of pigs and otherspecies of interest, especially cattle.

ACKNOWLEDGMENTS

We thank Mary Kenney and Mital Pandya (Plum Island AnimalDisease Center, ARS, USDA) for their support and assistance withthese studies. We also thank staff of the Animal Resources Unit (PlumIsland Animal Disease Center, DHS) for their professional work withthe animals used in these experiments.

This work was funded by CRIS grant no. 1940-32000-052-00D fromthe Agricultural Research Service, USDA (W.T.G.). In addition, thiswork was partially cofunded by the Danish Council for IndependentResearch, Technology and Production Sciences grant no. 274-09-0281(G.J.). L. E. Pedersen was the recipient of a Plum Island AnimalDisease Center Research Participation Program fellowship, adminis-tered by the Oak Ridge Institute for Science and Education (ORISE)through an interagency agreement between the U.S. Department ofEnergy (DOE) and the U.S. Department of Agriculture (USDA).

REFERENCES

1. Bautista, E. M., G. S. Ferman, and W. T. Golde. 2003. Induction of lym-phopenia and inhibition of T cell function during acute infection of swinewith foot and mouth disease virus (FMDV). Vet. Immunol. Immunopathol.92:61–73.

2. Belsham, G. J., and J. K. Brangwyn. 1990. A region of the 5� noncodingregion of foot-and-mouth disease virus RNA directs efficient internal initi-ation of protein synthesis within cells: involvement with the role of L pro-tease in translational control. J. Virol. 64:5389–5395.

3. Belsham, G. J., G. M. McInerney, and N. Ross-Smith. 2000. Foot-and-mouthdisease virus 3C protease induces cleavage of translation initiation factorseIF4A and eIF4G within infected cells. J. Virol. 74:272–280.

4. Carrillo, C., et al. 2005. Comparative genomics of foot-and-mouth diseasevirus. J. Virol. 79:6487–6504.

5. Chavez-Galan, L., M. C. Arenas-Del Angel, E. Zenteno, R. Chavez, and R.Lascurain. 2009. Cell death mechanisms induced by cytotoxic lymphocytes.Cell Mol. Immunol. 6:15–25.

6. Childerstone, A., L. Cedillo-Baron, M. Foster-Cuevas, and R. Parkhouse.1999. Demonstration of bovine CD8� T-cell responses to foot-and-mouthdisease virus. J. Gen. Virol. 80:663–669.

7. Chinsangaram, J., M. Koster, and M. J. Grubman. 2001. Inhibition ofL-deleted foot-and-mouth disease virus replication by alpha/beta interferoninvolves double-stranded RNA-dependent protein kinase. J. Virol. 75:5498–5503.

8. de Bruin, T. G., et al. 2000. Discrimination of different subsets of cytolyticcells in pseudorabies virus-immune and naive pigs. J. Gen. Virol. 81:1529–1537.

9. de Los Santos, T., S. de Avila Botton, R. Weiblen, and M. J. Grubman. 2006.The leader proteinase of foot-and-mouth disease virus inhibits the inductionof beta interferon mRNA and blocks the host innate immune response.J. Virol. 80:1906–1914.

10. de Los Santos, T., F. Diaz-San Segundo, and M. J. Grubman. 2007. Degra-dation of nuclear factor kappa B during foot-and-mouth disease virus infec-tion. J. Virol. 81:12803–12815.

11. Denyer, M. S., T. E. Wileman, C. M. Stirling, B. Zuber, and H. H. Taka-matsu. 2006. Perforin expression can define CD8 positive lymphocyte sub-sets in pigs allowing phenotypic and functional analysis of natural killer,cytotoxic T, natural killer T and MHC un-restricted cytotoxic T-cells. Vet.Immunol. Immunopathol. 110:279–292.

12. Devaney, M. A., V. N. Vakharia, R. E. Lloyd, E. Ehrenfeld, and M. J.Grubman. 1988. Leader protein of foot-and-mouth disease virus is requiredfor cleavage of the p220 component of the cap-binding protein complex.J. Virol. 62:4407–4409.

13. Diaz-San Segundo, F., et al. 2006. Selective lymphocyte depletion during theearly stage of the immune response to foot-and-mouth disease virus infectionin swine. J. Virol. 80:2369–2379.

14. Doel, T. R., L. Williams, and P. V. Barnett. 1994. Emergency vaccinationagainst foot-and-mouth disease: rate of development of immunity and itsimplications for the carrier state. Vaccine 12:592–600.

15. Falk, M. M., et al. 1990. Foot-and-mouth disease virus protease 3C inducesspecific proteolytic cleavage of host cell histone H3. J. Virol. 64:748–756.

16. Golde, W. T., C. K. Nfon, and F. N. Toka. 2008. Immune evasion duringfoot-and-mouth disease virus infection of swine. Immunol. Rev. 225:85–95.

17. Golde, W. T., et al. 2005. Vaccination against foot-and-mouth disease virusconfers complete clinical protection in 7 days and partial protection in 4days: use in emergency outbreak response. Vaccine 23:5775–5782.

18. Grubman, M. J., and B. Baxt. 2004. Foot-and-mouth disease. Clin. Micro-biol. Rev. 17:465–493.

19. Grubman, M. J., M. P. Moraes, F. Diaz-San Segundo, L. Pena, and T. de losSantos. 2008. Evading the host immune response: how foot-and-mouth dis-ease virus has become an effective pathogen. FEMS Immunol. Med. Micro-biol. 53:8–17.


20. Grubman, M. J., et al. 2010. Adenovirus serotype 5-vectored foot-and-mouthdisease subunit vaccines: the first decade. Future Virol. 5:51–64.

21. Grubman, M. J., M. Zellner, G. Bablanian, P. W. Mason, and M. E. Piccone.1995. Identification of the active-site residues of the 3C proteinase of foot-and-mouth disease virus. Virology 213:581–589.

22. Guzman, E., G. Taylor, B. Charleston, M. A. Skinner, and S. A. Ellis. 2008.An MHC-restricted CD8� T-cell response is induced in cattle by foot-and-mouth disease virus (FMDV) infection and also following vaccination withinactivated FMDV. J. Gen. Virol. 89:667–675.

23. Ho, C. S., M. H. Franzo-Romain, Y. J. Lee, J. H. Lee, and D. M. Smith. 2009.Sequence-based characterization of swine leucocyte antigen alleles in com-mercially available porcine cell lines. Int. J. Immunogenet. 36:231–234.

24. Ho, C. S., et al. 2009. Nomenclature for factors of the SLA system, update2008. Tissue Antigens 73:307–315.

25. Hoof, I., et al. 2009. NetMHCpan, a method for MHC class I bindingprediction beyond humans. Immunogenetics 61:1–13.

26. Joshi, G., R. Sharma, and N. K. Kakker. 2009. Phenotypic and functionalcharacterization of T-cells and in vitro replication of FMDV serotypes inbovine lymphocytes. Vaccine 27:6656–6661.

27. Kuhn, R., N. Luz, and E. Beck. 1990. Functional analysis of the internaltranslation initiation site of foot-and-mouth disease virus. J. Virol. 64:4625–4631.

28. Leisner, C., et al. 2008. One-pot, mix-and-read peptide-MHC tetramers.PLoS One 3:e1678.

29. Li, X., et al. 2008. Induction of protective immunity in swine by immuniza-tion with live attenuated recombinant pseudorabies virus expressing thecapsid precursor encoding regions of foot-and-mouth disease virus. Vaccine26:2714–2722.

30. Mayr, G. A., J. Chinsangaram, and M. J. Grubman. 1999. Development ofreplication-defective adenovirus serotype 5 containing the capsid and 3Cprotease coding regions of foot-and-mouth disease virus as a vaccine candi-date. Virology 263:496–506.

31. Mayr, G. A., V. O’Donnell, J. Chinsangaram, P. W. Mason, and M. J.Grubman. 2001. Immune responses and protection against foot-and-mouthdisease virus (FMDV) challenge in swine vaccinated with adenovirus-FMDVconstructs. Vaccine 19:2152–2162.

32. McCullough, K. C., et al. 1992. Relationship between the anti-FMD virusantibody reaction as measured by different assays, and protection in vivoagainst challenge infection. Vet. Microbiol. 30:99–112.

33. Moffat, K., et al. 2005. Effects of foot-and-mouth disease virus nonstructuralproteins on the structure and function of the early secretory pathway: 2BCbut not 3A blocks endoplasmic reticulum-to-Golgi transport. J. Virol. 79:4382–4395.

34. Moffat, K., et al. 2007. Inhibition of the secretory pathway by foot-and-mouthdisease virus 2BC protein is reproduced by coexpression of 2B with 2C, andthe site of inhibition is determined by the subcellular location of 2C. J. Virol.81:1129–1139.

35. Moraes, M. P., G. A. Mayr, P. W. Mason, and M. J. Grubman. 2002. Earlyprotection against homologous challenge after a single dose of replication-defective human adenovirus type 5 expressing capsid proteins of foot-and-mouth disease virus (FMDV) strain A24. Vaccine 20:1631–1639.

36. Nfon, C. K., G. S. Ferman, F. N. Toka, D. A. Gregg, and W. T. Golde. 2008.Interferon-alpha production by swine dendritic cells is inhibited during acuteinfection with foot-and-mouth disease virus. Viral Immunol. 21:68–77.

37. Nfon, C. K., F. N. Toka, M. Kenney, J. M. Pacheco, and W. T. Golde. 2010.Loss of plasmacytoid dendritic cell function coincides with lymphopenia andviremia during foot-and-mouth disease virus infection. Viral Immunol. 23:29–41.

38. Nielsen, M., et al. 2007. NetMHCpan, a method for quantitative predictionsof peptide binding to any HLA-A and -B locus protein of known sequence.PLoS One 2:e796.

39. Pacheco, J. M., M. C. Brum, M. P. Moraes, W. T. Golde, and M. J. Grubman.2005. Rapid protection of cattle from direct challenge with foot-and-mouthdisease virus (FMDV) by a single inoculation with an adenovirus-vectoredFMDV subunit vaccine. Virology 337:205–209.

40. Pacheco, J. M., et al. 2010. IgA antibody response of swine to foot-and-mouth disease virus infection and vaccination. Clin. Vaccine Immunol. 17:550–558.

41. Pauly, T., et al. 1995. Classical swine fever virus-specific cytotoxic T lympho-cytes and identification of a T cell epitope. J. Gen. Virol. 76(pt. 12):3039–3049.

42. Pauly, T., et al. 1996. Differentiation between MHC-restricted and non-MHC-restricted porcine cytolytic T lymphocytes. Immunology 88:238–246.

43. Pena, L., et al. 2008. Delivery of a foot-and-mouth disease virus empty capsidsubunit antigen with nonstructural protein 2B improves protection of swine.Vaccine 26:5689–5699.

44. Piriou-Guzylack, L., and H. Salmon. 2008. Membrane markers of the im-mune cells in swine: an update. Vet. Res. 39:54.

45. Robertson, B. H., D. O. Morgan, and D. M. Moore. 1984. Location ofneutralizing epitopes defined by monoclonal antibodies generated againstthe outer capsid polypeptide, VP1, of foot-and-mouth disease virus A12.Virus Res. 1:489–500.

46. Rock, K. L., D. J. Farfan-Arribas, and L. Shen. 2010. Proteases in MHC classI presentation and cross-presentation. J. Immunol. 184:9–15.

47. Rodriguez, A., et al. 1996. A porcine CD8� T cell clone with heterotypicspecificity for foot-and-mouth disease virus. J. Gen. Virol. 77(pt. 9):2089–2096.

48. Rodriguez, L. L., and M. J. Grubman. 2009. Foot and mouth disease virusvaccines. Vaccine 27(suppl. 4):D90–D94.

49. Saalmuller, A., T. Pauly, B. J. Hohlich, and E. Pfaff. 1999. Characterizationof porcine T lymphocytes and their immune response against viral antigens.J. Biotechnol. 73:223–233.

50. Sanz-Parra, A., et al. 1999. Recombinant viruses expressing the foot-and-mouth disease virus capsid precursor polypeptide (P1) induce cellular butnot humoral antiviral immunity and partial protection in pigs. Virology259:129–134.

51. Sanz-Parra, A., F. Sobrino, and V. Ley. 1998. Infection with foot-and-mouthdisease virus results in a rapid reduction of MHC class I surface expression.J. Gen. Virol. 79(pt. 3):433–436.

52. Sanz-Parra, A., et al. 1999. Evidence of partial protection against foot-and-mouth disease in cattle immunized with a recombinant adenovirus vectorexpressing the precursor polypeptide (P1) of foot-and-mouth disease viruscapsid proteins. J. Gen. Virol. 80(pt. 3):671–679.

53. Stanway, G., et al. 2005. Virus taxonomy: classification and nomenclature ofviruses: eighth report of the International Committee on the Taxonomy ofViruses. Elsevier Academic Press, San Diego, CA.

54. Tesar, M., and O. Marquardt. 1990. Foot-and-mouth disease virus protease3C inhibits cellular transcription and mediates cleavage of histone H3. Vi-rology 174:364–374.

55. Toka, F. N., C. Nfon, H. Dawson, and W. T. Golde. 2009. Natural killer celldysfunction during acute infection with foot-and-mouth disease virus. Clin.Vaccine Immunol. 16:1738–1749.

56. Toka, F. N., C. K. Nfon, H. Dawson, D. M. Estes, and W. T. Golde. 2009.Activation of porcine natural killer cells and lysis of foot-and-mouth diseasevirus infected cells. J. Interferon Cytokine Res. 29:179–192.

57. Yang, M., A. Clavijo, R. Suarez-Banmann, and R. Avalo. 2007. Productionand characterization of two serotype independent monoclonal antibodiesagainst foot-and-mouth disease virus. Vet. Immunol. Immunopathol. 115:126–134.

58. Yewdell, J. W., and S. M. Haeryfar. 2005. Understanding presentation ofviral antigens to CD8� T cells in vivo: the key to rational vaccine design.Annu. Rev. Immunol. 23:651–682.


45

7. SLA Tetramers

The identification of peptides strongly bound by SLA molecules, and producing stable pSLA

complexes, was performed in regard to (i) achieving knowledge about the individual SLA

binding profiles, and (ii) producing stable pSLA complex based tetramers for staining of porcine

PBMCs.

7.1. FMDV TETRAMER STAINING

pSLA tetramer studies were performed at PIADC as part of this PhD project using FMDV derived

peptides, followed by staining of PBMCs isolated from six animals previously vaccinated and

boosted against FMDV. These studies resulted in the identification of an FMDV T cell epitope

(Figure 13). General results and findings of the overall study are fully discussed in Paper IV, and

presents the first demonstration of FMDV-specific CTL killing of infected porcine kidney (PK15)

cells with confirmation by SLA allele matched tetramer staining.

My primary focus will be on the tetramer part of the studies presented in Paper IV, in that it

constitutes a general section of the overall PhD project. Hence, the killing analysis and results

are only briefly discussed.

To date the role of CTLs in FMDV immunity has gone largely unexplored partly due to the role

of neutralizing Abs been proven dominating in the establishment of protection against the virus

as described above. Furthermore, the ability of the virus to inhibit MHC synthesis and transport

to the surface has previously been described in vitro (Sanz-Parra et al. 1998). The inhibited MHC

synthesis and transport combined with the virus-induced lymphopenia, predict FMDV-specific

CTLs to play a minor role in clearing natural FMDV infection, if the MHC class I expression is as

well obstructed in vivo. Further complicating the analysis of specific CTLs and their role during

FMDV infection is that FMDV infected cells are rapidly lysed in vitro, and hence cannot be used

as targets in assays for antigen-specific cell lysis.

We developed an assay using available strains of pigs homozygous for SLA-1*0401, -2*0401,

and -3*0401. Further, the porcine fibroblast cell line, PK15, cells are a genetic match to the SLA

class I genes expressed in these pigs and thus are appropriate targets for the in vitro killing

46

assay. Using the replication defective Ad5 virus vector expressing FMDV P1 protein, we were

able to infect PK15 target cells without lysis and thus test CTLs from vaccinated, MHC matched

pigs for killing of the infected PK15 cells. A vaccine construct containing a mutation in the

protease 3Cpro (Ad5-FMDV-T) enabled us to test a hypothesis that an increase in the

intracellular pool of FMDV proteins available for degradation and loading of FMDV-derived

peptides into SLA class I molecules would enhance stimulation of FMDV peptide specific CTLs.

To confirm that the observed CTL induced killing was antigen specific, and MHC restricted,

FMDV pSLA complex based tetramers were produced and used to stain PBMCs isolated from

the vaccinated animals at identical time points post vaccine boost as for the PBMCs used to kill

infected PK15 targets. A comparison of animals vaccinated using the Ad5-FMDV-T vaccine with

animals vaccinated with a 3Cpro intact vaccine (Ad5-FMDV-B) was performed, to verify the

hypothesis of enhanced CTL stimulation due to inactive 3Cpro (Figure 13 A versus B, Day 4).

Figure 13. Tetramer staining of CD8+CD4

- PBMCs using two tetramers produced with either an FMDV (red) or Ebola

(blue) virus-derived peptide. Histograms of tetramer stained CD8+CD4

- PBMCs from Ad5-FMDV-B (A) or Ad5-

FMDV-T (B) –vaccinated animals are shown. Numbers indicate the individual percentage FMDV tetramer positive

cells within the CD8+CD4

- population of PBMCs after subtracting the percentage of cells stained with the negative-

control Ebola-tetramer. The figure shows responses for three Ad5-FMDV-B vaccinated pigs (p864, 865, 866) and

three Ad5-FMDV-T vaccinated pigs (p925, 945, 947). The figure is taken from Paper IV of this thesis.

47

As illustrated in Figure 13, animals vaccinated with the Ad5-FMDV-T construct stained clearly

positive with tetramer at day 4 post vaccine boost (Figure 13B), whereas the Ad5-FMDV-B

vaccinated animals showed no staining for samples taken at the same time point (Figure 13A).

Similar staining was seen for the Ad5-FMDV-B vaccinated group of animals pre-boost and on

day 8 (not shown). These data confirm the hypothesis stated above and classifies the Ad5-

FMDV-T construct as a CTL inducing vaccination approach against FMDV, and identify the FMDV

derived peptide of the positive tetramer (MTAHITVPY) as a CTL epitope capable of being

recognized by a significant number of activated CTLs within the CD8+CD4- population (ranging

from 5.63 – 20.42 % between different animals). An increase in tetramer staining in the scale of

2- to 3-fold was seen for Ad5-FMDV-T vaccinated animals on day 4 post boost as compared to

pre-boost levels (Figure 13B) and correlated with peak cytolysis of adenovirus infected PK15

target cells observed at day 4 post boost (Figure 6, Paper IV). In a previous study performed in

cattle, it was found that bovine leukocyte antigen (BoLA) class II tetramers (DRB3*1201-P16-7)

positively stained between 10 and 15.5 % of bovine CD4+ T cells derived from cattle previously

immunized once (Norimine et al. 2006). Instead of boosting their animals with antigen,

Norimine et al. used in vitro stimulation of positively selected CD4+ lymphocytes for one week

with peptide P16-7 to achieve proliferation of the peptide specific cell population. We observed

a decrease in cytolysis and absence of positive tetramer staining at day 8 post boost (Figure 13B

and Paper IV, Figure 6), which might be an indication of the dose and duration of the Ad5

antigen expression not being efficient enough to create a durable response for longer periods

without further boosting. A certain level of pre-boost staining was observed for p945 and p947

(Figure 13B) which was likely due to previous vaccination of the animals followed by a primary

boost 12 weeks post vaccination. The tetramer staining presented in Figure 13 was performed

following a second boost at 11 (FMDV-T) and 13 (FMDV-B) months post vaccination. Altogether

these data serve as confirmation of the presence of anti-FMDV, SLA-restricted CTLs and

illustrate the ability of the methods, developed in this project to fit the porcine model, in the

identification of CTL epitopes and monitoring of immune responses induced by vaccination.

48

7.2. INFLUENZA TETRAMER STAINING

Approaching the end of the PhD project, having established in-house SLA molecules with

several identified high affinity peptide binders derived from Influenza along with the tools

available to specifically monitor activated porcine CTLs, I got the opportunity to get fresh blood

samples from 20 animals which were part of another project. These animals had been

immunized five times against swine influenza A with three week intervals. Unfortunately, once I

got knowledge about their existence they had just received their fifth and final immunization

and faced elimination. Hence I only had one chance to try and identify any influenza epitopes.

The animals had received chemically (C3H4O2) inactivated TCID50/ml doses between 2.32 x 105

and 2.32 x 107 of various swine influenza A strains divided into different groups of pigs and

delivered intramuscularly in the ratio of 1:1 in Freunds Incomplete adjuvant (Table 2). Blood

samples were collected from all 20 pigs and each of them was SLA allele typed using the PCR-

SSP typing method and custom SLA class I designed high resolution primers. Of the 20 animals

typed, 16 turned out to express the allele of interest (SLA-1*0401) for which we had previously

identified four influenza peptides, CTELKLSDY, GTEKLTITY, SSSFSFGGF, and YVFVGTSRY, bound

with high affinities. These influenza peptides were predicted to be bound by NetMHCpan in

conjunction with the SLA-1*0401 binding matrix criteria, and identified as strongly bound

peptides by in vitro analysis in the same way as described for the FMDV derived peptides

presented in Paper II.

Table 2. Comparison of influenza peptide sequences within the different strains used for immunizations. Only

animals with cells found to positively stain with tetramer are shown (pigs 1, 2, 4, 6, 8 and 16) along with pig 17

which represents the 10 non-responding animals. (+) Peptide sequence present in the viral strain used for

immunization. (*) Position relative to start codon in virus A/swine/Denmark/12687/2003 (Breum 2012). (†)

Reassortant swine influenza virus encoding a different N2 gene (Breum 2012). (‡) Reassortant influenza virus strain

Tetramer peptides and positions Immunization strains 1 - 6

in Influenza viral genome 1 2 3 4 5 6

Peptide

Viral protein

of origin

AA position*

in virus

Nucleotide

position* in

virus

A/swine/Den

mark/101310-

1/2011

(H1N1pdm09)

A/swine/Den

mark/101568-

1/2011

(H1pdm09†)

A/swine/Den

mark/19126/

1993 (H1N1)

A/swine/Den

mark/101490-

3/2011

(H1N1)

A/swine/Den

mark/1037-

2/2011

(H1N2†)

A/swine/Den

mark/101116-

3/2011

(H1N2‡)

CTELKLSDY NP 44 - 52 130-156 + + CTELQLSDY CTELQLSDY CTELQLSDY +

GTEKLTITY PB2 523 - 531 1567-1593 + + + + + +

SSSFSFGGF PB2 320 - 328 958-984 + + + + + +

YVFVGTSRY HA 215 - 223 643-669 + + YVSVESSKY YVSVVSSKY YVSVVSSKY YVSVGSSKY

Animal ID for pigs immunized by the respective viral strains 1 - 6

1 + 2 4 6 8 15 + 16 17

49

with NP gene from H1N1pdm09 virus (Breum 2012). Amino acids in bold (grey), mark substitutions in the peptide

sequence coded for in the immunization strain as compared to the peptide used for tetramer production.

PBMCs isolated from the 16 animals positive for SLA-1*0401 were initially frozen and later

analyzed by influenza pSLA-1*0401 complex based tetramer staining using four different

influenza peptides and two different fluorochrome tetramer labels (allophycocyanin (APC) and

brilliant violet (BV421)), which enable the identification of tetramer double positive cell

populations as illustrated in Figure 14A (Pig 4, (GTEKLTITY)). A negative control peptide

(ASYGAGAGY) was included for comparison. The experimental setup for these immunizations

was designed for other purposes than tetramer staining, which left us with no non-vaccinated

or SLA mismatched pigs as negative controls. Hence, I decided to use a human PBMC sample

stained with tetramers as an additional negative control. Such a control has to be included with

caution in that it is not a definite negative control. Still, it is capable of proving that the

tetramers we use are not just staining whatever cells they are offered regardless of their

species of origin.

Of the 16 correctly SLA matched animals it was found that six pigs (Pig 1, 2, 4, 6, 8, and 16)

displayed positive signals above background for one or more of the influenza tetramers in the

staining assay (Table 3). A study examining influenza virus specific CD8+ T cell responses in

Macaques has recently been performed using macaque MHC class I, Mane-A1*08401,

tetramers produced with an influenza epitope; RA9 (Jegaskanda et al. 2012). In this study

animals were infected intranasal and intratracheal at day 0 and day 28, followed by

administration of a third infection dose 2 to 10 months later using recombinant influenza virus.

PBMCs were collected from blood and frozen. The results of these experiments showed distinct

staining of thawed RA9-specific CD8+ T cells by 0.14 % of the CD8+ T cell population compared

to 0.02 % negative control staining of an influenza naïve Mane-A1*08401-RA9 negative animal,

corresponding to a 7-fold increase in tetramer staining signal between positive and negative

samples.

From our studies, tetramer cell staining examples of double positively stained (SLA matched,

responding animal 4), no staining (SLA matched, non-responding animal 17) and additional

negative control staining of human PBMCs with SLA tetramers (SLA mismatched, non-

50

responding human) are shown in Figure 14 for one swine influenza A peptide (GTEKLTITY) and

the negative control peptide (ASYGAGAGY). Here illustrating a 5.4-fold increase in tetramer

staining signal between positive and negative samples (Figure 14A). A 32.5-fold increase in

positive tetramer staining was seen for PBMCs derived from pig 4 and stained with the

influenza tetramer (GTEKLTITY) compared with the same tetramer staining of completely

mismatched human PBMCs (Figure 14A + C). Comparing the influenza-tetramer data for the six

pigs with those of the macaque study performed by Jegaskanda et al., it is demonstrated that

our positive tetramer staining percentages are generally 11 to 46-folds higher, but with similar

ratios to the negative control samples, and with some variation between different animals and

peptides. These tetramer staining percentages are generally higher than previously observed

for fresh ex vivo staining of CD8+ PBMCs responding to vaccination against (or infection with)

yellow fever (Co et al. 2009) in humans, M. tuberculosis (Wei et al. 2009) in macaque monkeys,

and hepatitis C (He et al. 1999) in humans, and could be due to a higher degree of auto-

fluorescent cells when studying porcine PBMCs or as a result of a higher level of non-specific,

false positive tetramer background staining using SLA tetramers. Most of the responses

reported in humans and macaques were shown to increase in tetramer positive percentage

scores once stimulated with antigen, hereby reaching numbers around (or above) what is

reported here for staining of porcine PBMCs.

Neg. Ctrl. Peptide

Pig 4 (ASYGAGAGY)

Tetramer BV-A

Te

tram

er

AP

C-A

Tetramer BV-A

Tetr

am

er

AP

C-A1.2 % 6.5 %

Influenza A peptide

Pig 4 (GTEKLTITY)

A

51

Figure 14. Tetramer staining of PBMCs. (A + B): PBMCs from animals previously vaccinated and boosted against

swine influenza A. (C): Human PBMCs. Individual samples were stained by a negative control peptide (ASYGAGAGY)

tetramer and a swine influenza A derived peptide (GTEKLTITY) tetramer. Singlet PBMCs were gated in P1 (blue).

CD8+(high)CD4

- cells were gated in P2 (orange), and CD8

+(high)CD4

-APC

+BV

+ double tetramer positive cells were

gated in P3 (green) for Pig 4 and Pig 17 (A + B). For the human control samples, singlet PBMCs were gated in P1

(blue), CD8+(high) cells were gated in P2 (orange) and CD8

+(high)BV

+ tetramer single positive cells were gated in P3

(green) (C). Percentages of tetramer positive cells within the CD8+(high)CD4

- population are shown for each sample

in the figure. Individual staining percentages for all animals and tetramers are given in Table 3.

Tetramer BV-A

Te

tram

er

AP

C-A

Tetramer BV-A

Te

tram

er

AP

C-A

Neg. Ctrl. Peptide

Pig 17 (ASYGAGAGY)

Influenza A peptide

Pig 17 (GTEKLTITY)

0.3 % 0.5 %

B

Neg. Ctrl. peptide

Human (ASYGAGAGY)

Influenza A peptide

Human (GTEKLTITY)

Tetramer BV-A

Tetramer BV-A

0.2 %

0.2 %

C

52

The influenza tetramer studies presented here revealed four influenza derived peptide epitopes

(CTELKLSDY, GTEKLTITY, SSSFSFGGF, YVFVGTSRY) which were recognized in complex with the

SLA-1*0401 molecule by porcine CTLs activated by immunization against swine influenza as

illustrated in Figure 15 for Pig 4. Not all animals were responsive towards every influenza

tetramer as positive staining was only seen for one of the four tetramers for pigs 1 and 2 (Table

3) using a threshold of 1 % positive staining of the CD8hi+ PBMC population post negative

background subtraction, which corresponds to a 2-fold higher staining percentage than

measured for the negative control tetramer stained cells. Such a 2-fold higher threshold has

previously been set by others (Ferrari et al. 2004). Furthermore, it was found that 10 of the 16

animals did not stain positive with any of the tetramers, as illustrated by Pig 17 (Figure 14B +

15, and Table 3), despite being correctly SLA matched. Such non-responsiveness illustrates how

SLA bound peptides are not necessarily recognized as epitopes in all animals. Similar

observations have previously been made for HLA tetramers and yellow fever peptides as well

(Stryhn 2012). Collectively, the differences in individual peptide responses of animals with

identical SLA alleles indicate that the SLA-1*0401 tetramers were not just staining any activated

CTL derived from a correctly matched animal, but only CTLs with the right specificity, and that

although peptide binding to the SLA molecule may be predicted and verified in advance, this is

no guarantee for development of a CTL specific immune response to this particular peptide.

53

Figure 15. Percentages of tetramer positive cells within the CD8+(high)CD4

- population stained by each of the four

influenza tetramers. Cell populations and individual percentages are shown in Figure 14 and Table 3, respectively.

All four influenza tetramer peptides tested, along with the negative control tetramer (ASYGAGAGY), are shown

(black and grey bars, respectively).

An interesting observation was made for Pigs 6, 8 and 16 which had all received viral

immunization strains that differ in the NP and HA protein regions coding for two of the peptide

sequences used to produce the tetramers (Table 2). The peptide CTELKLSDY had a single

substitution in P5 of the peptide (K48Q) compared to the viral strains used for immunization of

animals 6, 8 and 16, whereas peptide YVFVGTSRY had up to four substitutions (F217S, G219E,

T220S, R222K) for peptide residues P3, P5, P6, and P8. Still, positive tetramer staining of CTLs

was seen for all three pigs using these peptides. Such observations could be explained by CTL

cross-reactivity suggesting that CTL clones can recognize different peptides, some even sharing

no obvious sequence identity (Frankild et al. 2008). In addition to this, looking at the nature of

the AAs involved in the substitutions could explain why the CTELKLSDY based tetramers were

recognized by the porcine CTLs in that lysine (K) features a basic, positively charged -NH3+ side

chain and glutamine (Q) features a neutral amide -CONH2 side chain, making both AAs

hydrophilic and capable of protrude toward the solvent. From a peptide:SLA-1*0401 affinity

54

point of view it is also seen that both K and Q are bound with almost similar RB values of 0.4

and 0.5, respectively (Paper I, Table 1). For the tetramers produced with the YVFVGTSRY

peptide, the positive staining could be reasoned by similarities between the different AA

residues responsible for the substitutions and the respective AA residues within the peptides

derived from the viral strains. According to the SLA-1*0401 binding matrix it can be seen that

no major differences exist between the different AA substitutions when it comes to binding

affinity RB values, slightly favoring the viral strain derived peptide sequences (YVSVESSKY,

YVSVVSSKY, YVSVGSSKY) for binding compared to the tetramer peptide sequence (Paper I,

Table 1). Both phenylalanine (F) and serine (S) are being hidden within the D-pocket of the SLA-

1*0401 peptide binding groove responsible for binding residue P3, with neither of them being

disfavored for binding according to the binding matrix (Paper I, Table 1). Hence, despite their

overall structural diversity, it could be speculated that such substitution plays a minor role in

the matter of tetramer staining, in that these residues are not available for recognition by the

TCR. Furthermore, the only difference seen between threonine (T) and serine in P6 is the

methyl group of threonine, seeing both AAs as nucleophilic and hydrophilic which makes them

more or less alike in terms of interacting with the surroundings. For residue P8 the AAs arginine

(R) and lysine (K) share common features in being positively charged, basic and hydrophilic

enabling almost similar recognition, whereas the AAs in P5 truly differ in glycine (G) being small

and hydrophobic and glutamic acid (E) is acidic and hence hydrophilic. Comparing the other P5

substitution of (G219V) it is noteworthy that both AAs are hydrophobic, with valine (V) being

somewhat larger in the presentation of a dimethyl group to the solvent. These P5 AA residues

point away from the groove of the MHC enabling them to interact with the TCR of specific CTLs.

When that is said, glycine being the smallest existing AA available might enable TCR recognition

of the YVFVGTSRY-based tetramers, in that it is likely that the glycine P5 residue does not play a

significant role in this particular TCR:pSLA interaction, or at least is unlikely to interfere with

such. Nevertheless, further experiments including further negative controls such as non-

vaccinated and SLA mismatched animals, as well as pre-vaccination and pre-boost samples for

comparison, would be needed for conclusively confirmation of the specific influenza tetramer

staining data presented in this section.

55

Table 3. Specific staining of porcine CD8+(high)CD4

- PBMCs using four different influenza peptide-based tetramers.

Percentages in bold illustrate positive tetramer staining based on a threshold of 1.0 % post background

subtraction. Italic numbers illustrate no specific binding whereas grey highlights point out positive staining with

peptides that are not 100 % identical in sequence with those of the respective vaccine constructs used.

ASYGAGAGY is a designed negative control peptide for defining the relative false-positive background signal.

Animal ID 17 was included as an example of a non-responding animal. ID Human: Human PBMCs included as an

additional negative control for comparison.

8. Conclusions and Perspectives

Throughout this PhD project, the fully developed large-scale technology to produce, analyze,

and validate peptide–MHC interactions, which was pioneered over three decades for mice and

humans (Ferre et al. 2003, Harndahl et al. 2009, Ostergaard et al. 2001, Pedersen et al. 1995,

Pig #

Tetramer

influenza

peptide

Peptide present in

immunization strain used

% Tetramer+ cells

within the

CD8+CD4-

population

% Tetramer+ cells within

the CD8+CD4- population

(background subtracted)

ASYGAGAGY No 0,80 0,00

CTELKLSDY Yes 1,70 0,90

1 GTEKLTITY Yes 1,90 1,10

SSSFSFGGF Yes 1,70 0,90

YVFVGTSRY Yes 1,60 0,80












CTELKLSDY No 5,80 3,20



YVFVGTSRY No 5,90 3,30
















ASYGAGAGY - 0,20 0,00

CTELKLSDY - 0,30 0,10

Human GTEKLTITY - 0,20 0,00

SSSFSFGGF - 0,30 0,10

YVFVGTSRY - 0,30 0,10

56

Stryhn et al. 1996), has been applied to analysis of porcine MHC molecules. By doing so,

instigation of knowledge on potential T cell epitopes in this particular species, has been made

possible further allowing in silico prediction servers, that identify specific peptides for MHC

binding, such as the NetMHCpan, also to successfully include swine (Hoof et al. 2009, Larsen et

al. 2005, Lundegaard et al. 2011, Nielsen et al. 2007, Pedersen et al. 2011, Pedersen et al.

2012a). Combined with the tetramer technology described in this thesis for specific T cell

analysis this has enabled fast, accurate “one-pot, mix-and-read” analyses of MHC restricted

CD8+ T cell reactivity following vaccination. In pigs, in silico prediction of virus derived peptides

bound by SLA class I molecules has been verified by in vitro analysis of specific peptide binding,

leading to the identification of strongly bound viral peptides (Pedersen et al. 2011, Pedersen et

al. 2012a). In addition, the analysis of pSLA stability was included hereby identifying peptides,

which were not only bound with high affinity but further produce pSLA complexes of high

stability (Pedersen et al. 2012b). Furthermore, the effects of T cell specific vaccination have

been investigated by monitoring CD8+ T cell responses post vaccination (Patch et al. 2011). Such

reverse immunology proposals should be expected to become important in the contribution to,

as well as support of, the study of livestock immune responses.

As part of the project, SLA molecules have been produced and their peptide-binding profiles

analyzed. All SLA alleles were initially identified by PCR-SSP allele specific tissue typing (or

previously reported in the literature) to be commonly expressed among swine herds. Following

these approaches, I predicted and identified several peptides which were bound with high

affinity and generate highly stable pSLA complexes with one or more of the newly produced

SLA proteins. This was done by using the NetMHCpan prediction tool in concert with individual

SLA peptide binding matrix predictions, and followed by in vitro peptide affinity and stability

measures (Hoof et al. 2009, Pedersen et al. 2011, Pedersen et al. 2012b, Pedersen et al. 2012a).

Having identified such possible peptide epitopes led to SLA based tetramer staining

experiments in an attempt to positively stain PBMCs from animals previously vaccinated against

either FMDV (Patch et al. 2011) or swine influenza A. These experiments ultimately led to the

identification of T cell epitopes as presented in this thesis (Figure 13 and 15).

Altogether, the assays and experimental approaches presented and discussed here provide new

tools to (i) identify commonly expressed SLA class I alleles within herds, (ii) analyze the diverse

57

peptide binding profiles of individual SLA proteins, and (iii) identify antigenic peptide ligands

bound with high affinity and stability by such MHC molecules of interest. These results lead to

the production of pSLA tetramers, which, when combined with cytokine release and staining

assays, as well as measuring target cell killing, are capable of (i) providing a tool for the

enumeration and determination of the relative frequencies of epitope-specific, cytokine-

secreting CTLs in a sample, (ii) evaluating the effector function of such CTLs, i.e. their ability to

specifically kill infected targets and the identification of specific cytokines and chemokines

produced, and (iii) identifying viral as well as tumor derived peptide T cell epitopes (Patch et al.

2011, Romero et al. 1998, Romero et al. 2004).

Some future perspectives and suggested ideas for new projects based on the assays,

approaches and results presented in this thesis are illuminated in the sub sections 8.1. – 8.3.

8.1. SIMULTANEOUS ANALYSIS OF MULTIPLE SPECIFICITIES

An interesting approach for future studies would be the establishment of assays capable of

simultaneously analyzing porcine CTLs with different specificities based on differently labeled

pMHC complex multimers such as the systemically organized dextramers. Dextramers are

fluorescently labeled MHC class I oligomers based on dextran backbones, which are capable of

binding ten times more biotinylated pMHC complexes, and the conjugation of four to five times

more fluorescein than the conventional tetramers (Batard et al. 2006). They have been shown

to produce similar or stronger signal than their tetramer counterparts partly due to stronger

pMHC:TCR avidity, and particularly because of the higher level of fluorescent labeling. Hence, in

contrast to tetramers, MHC dextramers are not restricted to only the brightest fluorochromes

for optimal detection. Such features make dextramers interesting alternatives in the

identification and labeling of minor populations of antigen specific CTLs using higher order

multicolor flow cytometry, and enabling the use of a broad range of fluorochrome conjugates

including the ones that are less fluorescent (Batard et al. 2006). Using multicolor staining with

either dextramers or tetramers would further improve the cell staining assays and approaches

described in this thesis. Such multicolor staining could enable simultaneous identification and

enumeration of separate cell populations specific for individual pMHC complexes even after

58

surface-marker gating. This could be done using two different pMHC multimers each with their

own fluorochrome conjugate. In such ways, single (and dual) specificities of activated CTLs can

be demonstrated and the cell staining assay itself validated. The scenario of having dual

specificities expressed simultaneously by a single population of CTLs would be expected to

require almost identical pMHC complex based tetramers or cross-reactive CTLs. Applying such

pSLA tetramer or dextramer based multicolor staining assays to swine prove promising for

future studies of livestock diseases, and could turn out to become valuable tools, eventually

leading to advanced and high specific validation of the functions of existing vaccines, as well as

accelerated development of future vaccines.

8.2. SYNTHETIC IMMUNOGENIC PROTEIN (SIP) VACCINES

By use of the different assays and techniques presented throughout this thesis, it should be

possible to deliver a new and innovative approach for tailoring synthetic protein vaccines to

specifically induce an efficient CTL immune response against some of the most important

infectious diseases currently known to threat the swine production worldwide. Evaluation of

the different approaches to induce CTL responses with currently available adjuvants would be

facilitated by tetramer assessment of different vaccine formulations. An approach described

for, and tested in, mice more than a decade ago (Suhrbier 1997, Thomson et al. 1998) and later

followed up in regard to enhanced CTL responses using supporting adjuvant (Li et al. 2005),

make it reasonable to suggest novel vaccines can be designed of a synthetic string of conserved

sub-dominant, well-defined SLA class I CTL epitopes of viruses such as FMDV, PRRSV and swine

influenza A virus. Furthermore, it could be suggested that such a vaccine delivered either in a

CD8 activating adjuvant or in a live non-replicating vector, would be able to induce an efficient

and sustained CTL response against current viral strains as well as forthcoming virus mutations.

As described above, the recently developed SLA tetramers could be used to identify important

peptide epitopes (such as those highly conserved among viral strains) capable of activating

CTLs, followed by transformation of such identified peptides into a designed string of tandem

repeated peptides in a synthetic immunogenic protein (SIP) encoded by a carefully designed

DNA construct. Once the SIP is delivered, degraded and finally presented as peptides by antigen

presenting cells, the porcine adaptive immune system of vaccinated animals should become

59

activated in a highly specific manner leading to generation of immunologic memory and CTL

mediated protection against the virus represented by the SIP (Figure 16). Such SIP vaccines

could turn out to be a novelty with the potential to revolutionize current protein and subunit-

based vaccine approaches and development.

Figure 16. From genome to SIP vaccine to protection. 1: Selected genomic regions of the virus of interest is

screened for peptide candidates with the ability to form complexes with SLA class I molecules using combined in

silico NetMHCpan and PSCPL matrix predictions followed by in vitro verification. 2: Production of correctly folded

pSLA tetramers and the identification of T cell epitope peptides capable of being recognized by cytotoxic T cells. 3:

Design of DNA construct encoding tandem repeats of sequences encoding all identified viral epitopes. 4:

Recombinant production of the SIP directly from the DNA construct. 5: SIP-based vaccine formulation and adjuvant

non-vectored/vectored vaccine delivery of SIP proteins/construct. 6: Intracellular degradation of SIPs into peptides

and SIP peptide load onto SLA class I molecules, followed by cell surface presentation of SIP-pSLA complexes to

circulating CTLs ultimately leading to adaptive immune activation and protection against the virus of interest.

Investigation of the induction of CTL responses following live non-replicating adenovirus

vaccine vectored SIP (Ad-SIP) vaccination, or vaccination with the SIP formulated into different

adjuvants would be highly relevant in the search for optimal adjuvant-based co-delivery and

immune activation. Such novel adjuvants, with the potential to induce CTL responses, could

arise from modifications of the cationic adjuvant system (CAF) adjuvant family (Nordly et al.

2011), immune stimulatory complexes (ISCOMs) (Schiott et al. 2011), Cholera toxin (Olvera-

Gomez et al. 2012) or other new formulations (Heegaard et al. 2011). Succeeding in designing

60

functionally SIP vaccines would tremendously accelerate vaccine development against viral

infections which require a high quality host CTL response for elimination and establishment of

protection.

A potential drawback of the SIP based vaccination approach would be the challenge of getting

whole SIPs successfully delivered into the cytoplasm of antigen presenting cells, which make

protein-adjuvant vaccination a substantial task to deal with. This may be overcome by using the

Ad-SIP construct for the synthesis and delivery of SIPs inside the cells. Still, if the use of

adenovirus vectors is to be avoided, another solution could be adding “opsonization tags” in

the construct encoding the SIP, leading to specific glycosylation of the SIP, for example with

mannose, making it recognizable by mannose-binding lectins at the surface of antigen

presenting cells.

Most pork producers are likely to select breeding stock for quality and quantity of meat. Hence,

vaccine makers will have to customize vaccines to the strains of pigs favored by the producers.

If vaccines based on the most frequent SLA alleles are successful, however, it is likely that they

will also introduce selection towards an undesired reduction in SLA diversity. Thus, it will be

important to include a broad range of SLA specificities in the design of corresponding vaccine

polyepitopes.

8.3. MULTIMER BASED CANCER TREATMENT

Using swine as an animal model, it would be possible to validate the effects of human cancer

vaccines, by monitoring the specificity and status of activation of different cell populations via

pSLA multimers based on peptides derived from the cancer type of interest. Furthermore,

pMHC multimers could potentially be used in vivo to label tumor cells for elimination by already

existing CTLs of the host specific to common viruses such as Epstein-Barr virus (EBV),

Cytomegalo virus (CMV), and influenza, from which the tetramer peptides should be derived.

Hereby tetramers would resemble antigen presenting cells bound to the cell surface of tumor

cells by an anti-tumor specific mAb directly coupled to streptavidin of the tetramer (Figure 17),

enabling the activation of host CTLs and leading to elimination of the tumor cell.

61

Figure 17. Redirected killing of a tumor cell by an EBV specific CTL. EBV peptides (red) are presented by a pMHC

tetramer (green/purple), which is attached to the cell surface of a tumor cell (black) via an anti-tumor mAb (pink)

that recognize tumor antigens (grey).

In vitro studies have shown such Ab-pMHC tetramers to sensitize tumor cells to lysis by specific

CTLs (Ogg et al. 2000, Robert et al. 2000) making further refinement and development of this

technique for in vivo utilization an interesting approach for future immunological therapy of

cancer.

8.4. ANSWERS LEAD TO QUESTIONS

The approaches, assays and proposals presented in this PhD thesis could ultimately lead to a

better agriculture in terms of a more productive farming industry and overall greater animal

health. Still the questions are many, and the greatly accelerated evolution in the field of

immunology within the past two decades, has now made curiosity-driven research possible at a

scale not previously imaginable. It is often so in science that most answers tend to bring several

new questions. Hence, knowing what we do today forces us to strive for more. Could the

distribution of SLA molecules, such as the commonly expressed SLA-1*0401, be of any interest

in the susceptibility to, or protection against, certain viruses in swine? Would one commonly

62

occurring class I molecule cover for most of the pSLA complex mediated CTL immune response,

or would a symphony of several more or less common alleles be a criteria for the induction of

strong and durable adaptive immune protection? How important are the individual peptide

anchor requirements and peptide lengths for SLA class I ligand binding, and the corresponding

level of immunogenicity, and how does pSLA complex stability affect TCR activation? Would

high peptide turn-overs and a great variety in peptide presentation be beneficial to effective

protection against disease, or would a low number of a few highly stable pSLA complexes be

more capable of sufficiently activating CTLs? Would it be possible to specifically monitor

immune responses in outbred populations, leading to the identification of several conserved

peptide epitopes from viruses that are constantly mutating to achieve immune evasion? Could

such epitopes turn out to be potentially valuable vaccine candidates in the design and

development of highly specific subunit vaccines, and how should such a vaccine be delivered

for optimal effect? Would it be possible to effectively validate different vaccines and vaccine

delivery approaches for humans, using the pig as an experimental model, hereby improving our

understanding of targeting various cancers as well as of the pathogenesis and immune

elimination of disease in both humans and livestock animals?

Final Thoughts

To get a full – or at least better – understanding on what is going on in our field of veterinary

immunology, we need to step back and take yet another look, share and use the technologies

and approaches available to us in research laboratories around the globe; discard a lot, re-think

some, use a few, and finally have the courage to go down different alleys embracing whatever

complexities that we might find.

63

Reference List

Addo M.M., Yu X.G., Rathod A., Cohen D., Eldridge R.L., Strick D., Johnston M.N., Corcoran C., Wurcel

A.G., Fitzpatrick C.A., Feeney M.E., Rodriguez W.R., Basgoz N., Draenert R., Stone D.R., Brander C.,

Goulder P.J., Rosenberg E.S., Altfeld M. & Walker B.D. (2003) Comprehensive epitope analysis of human

immunodeficiency virus type 1 (HIV-1)-specific T-cell responses directed against the entire expressed

HIV-1 genome demonstrate broadly directed responses, but no correlation to viral load. J. Virol. 77,

2081-2092.

Altman J.D., Davis M.M. (2003) MHC-peptide tetramers to visualize antigen-specific T cells. Curr. Protoc.

Immunol. Chapter 17, Unit.

Altman J.D., Moss P.A., Goulder P.J., Barouch D.H., Heyzer-Williams M.G., Bell J.I., McMichael A.J. &

Davis M.M. (1996) Phenotypic analysis of antigen-specific T lymphocytes. Science 274, 94-96.

Anderson G., Takahama Y. (2012) Thymic epithelial cells: working class heroes for T cell development

and repertoire selection. Trends Immunol. 33, 256-263.

Banchereau J., Pascual V. & O'Garra A. (2012) From IL-2 to IL-37: the expanding spectrum of anti-

inflammatory cytokines. Nat. Immunol. 13, 925-931.

Baranov O.K., Eppel' M.S. (1984) [Genetic bases of antibody diversity]. Genetika 20, 1397-1413.

Batard P., Peterson D.A., Devevre E., Guillaume P., Cerottini J.C., Rimoldi D., Speiser D.E., Winther L. &

Romero P. (2006) Dextramers: new generation of fluorescent MHC class I/peptide multimers for

visualization of antigen-specific CD8+ T cells. J. Immunol. Methods 310, 136-148.

Beigel J., Bray M. (2008) Current and future antiviral therapy of severe seasonal and avian influenza.

Antiviral Res. 78, 91-102.

Bevan M.J. (1977) In a radiation chimaera, host H-2 antigens determine immune responsiveness of

donor cytotoxic cells. Nature 269, 417-418.

Bevan M.J. (1995) Antigen presentation to cytotoxic T lymphocytes in vivo. J. Exp. Med. 182, 639-641.

Bjorkman P.J., Saper M.A., Samraoui B., Bennett W.S., Strominger J.L. & Wiley D.C. (1987a) Structure of

the human class I histocompatibility antigen, HLA-A2. Nature 329, 506-512.

Bjorkman P.J., Saper M.A., Samraoui B., Bennett W.S., Strominger J.L. & Wiley D.C. (1987b) The foreign

antigen binding site and T cell recognition regions of class I histocompatibility antigens. Nature 329, 512-

518.

Breum S.Ø. Novel Influenza A viral strains and sequence analysis. 20-9-2012.

Ref Type: Personal Communication

Brocchi E., Bergmann I.E., Dekker A., Paton D.J., Sammin D.J., Greiner M., Grazioli S., De S.F., Yadin H.,

Haas B., Bulut N., Malirat V., Neitzert E., Goris N., Parida S., Sorensen K. & de C.K. (2006) Comparative

evaluation of six ELISAs for the detection of antibodies to the non-structural proteins of foot-and-mouth

disease virus. Vaccine 24, 6966-6979.

64

Brown I.H., Ludwig S., Olsen C.W., Hannoun C., Scholtissek C., Hinshaw V.S., Harris P.A., McCauley J.W.,

Strong I. & Alexander D.J. (1997) Antigenic and genetic analyses of H1N1 influenza A viruses from

European pigs. J. Gen. Virol. 78 ( Pt 3), 553-562.

Busch D.H., Pamer E.G. (1998) MHC class I/peptide stability: implications for immunodominance, in vitro

proliferation, and diversity of responding CTL. J. Immunol. 160, 4441-4448.

Campanini G., Piralla A., Paolucci S., Rovida F., Percivalle E., Maga G. & Baldanti F. (2010) Genetic

divergence of influenza A NS1 gene in pandemic 2009 H1N1 isolates with respect to H1N1 and H3N2

isolates from previous seasonal epidemics. Virol. J. 7, 209.

Chardon P., Renard C., Gaillard C.R. & Vaiman M. (2000) The porcine major histocompatibility complex

and related paralogous regions: a review. Genet. Sel Evol. 32, 109-128.

Chardon P., Rogel-Gaillard C., Cattolico L., Duprat S., Vaiman M. & Renard C. (2001) Sequence of the

swine major histocompatibility complex region containing all non-classical class I genes. Tissue Antigens

57, 55-65.

Choi E.M., Palmowski M., Chen J. & Cerundolo V. (2002) The use of chimeric A2K(b) tetramers to

monitor HLA A2 immune responses in HLA A2 transgenic mice. J. Immunol. Methods 268, 35-41.

Clavijo A., Wright P. & Kitching P. (2004) Developments in diagnostic techniques for differentiating

infection from vaccination in foot-and-mouth disease. Vet. J. 167, 9-22.

Co M.D., Kilpatrick E.D. & Rothman A.L. (2009) Dynamics of the CD8 T-cell response following yellow

fever virus 17D immunization. Immunology 128, e718-e727.

Cobbold M., Khan N., Pourgheysari B., Tauro S., McDonald D., Osman H., Assenmacher M., Billingham L.,

Steward C., Crawley C., Olavarria E., Goldman J., Chakraverty R., Mahendra P., Craddock C. & Moss P.A.

(2005) Adoptive transfer of cytomegalovirus-specific CTL to stem cell transplant patients after selection

by HLA-peptide tetramers. J. Exp. Med. 202, 379-386.

Cresswell P., Bangia N., Dick T. & Diedrich G. (1999) The nature of the MHC class I peptide loading

complex. Immunol. Rev. 172, 21-28.

Cresswell P., Turner M.J. & Strominger J.L. (1973) Papain-solubilized HL-A antigens from cultured human

lymphocytes contain two peptide fragments. Proc. Natl. Acad. Sci. U. S. A 70, 1603-1607.

Dalchau N., Phillips A., Goldstein L.D., Howarth M., Cardelli L., Emmott S., Elliott T. & Werner J.M. (2011)

A peptide filtering relation quantifies MHC class I peptide optimization. PLoS. Comput. Biol. 7, e1002144.

Danilova N. (2012) The evolution of adaptive immunity. Adv. Exp. Med. Biol. 738, 218-235.

Davis M.M., Bjorkman P.J. (1988) T-cell antigen receptor genes and T-cell recognition. Nature 334, 395-

402.

Doel T.R. (2003) FMD vaccines. Virus Res. 91, 81-99.

Doherty P.C., Zinkernagel R.M. (1975) H-2 compatibility is required for T-cell-mediated lysis of target

cells infected with lymphocytic choriomeningitis virus. J. Exp. Med. 141, 502-507.

65

Dunbar P.R., Ogg G.S., Chen J., Rust N., van der B.P. & Cerundolo V. (1998) Direct isolation, phenotyping

and cloning of low-frequency antigen-specific cytotoxic T lymphocytes from peripheral blood. Curr. Biol.

8, 413-416.

Elliott T., Cerundolo V., Elvin J. & Townsend A. (1991) Peptide-induced conformational change of the

class I heavy chain. Nature 351, 402-406.

Elliott T., Williams A. (2005) The optimization of peptide cargo bound to MHC class I molecules by the

peptide-loading complex. Immunol. Rev. 207, 89-99.

Engvall E., Perlmann P. (1971) Enzyme-linked immunosorbent assay (ELISA). Quantitative assay of

immunoglobulin G. Immunochemistry. 8, 871-874.

Essler S.E., Ertl W., Deutsch J., Ruetgen B.C., Groiss S., Stadler M., Wysoudil B., Gerner W., Ho C.S. &

Saalmueller A. (2012) Molecular characterization of swine leukocyte antigen gene diversity in purebred

Pietrain pigs. Anim Genet.

Ferrari G., Neal W., Ottinger J., Jones A.M., Edwards B.H., Goepfert P., Betts M.R., Koup R.A., Buchbinder

S., McElrath M.J., Tartaglia J. & Weinhold K.J. (2004) Absence of immunodominant anti-Gag p17 (SL9)

responses among Gag CTL-positive, HIV-uninfected vaccine recipients expressing the HLA-A*0201 allele.

J. Immunol. 173, 2126-2133.

Ferre H., Ruffet E., Blicher T., Sylvester-Hvid C., Nielsen L.L., Hobley T.J., Thomas O.R. & Buus S. (2003)

Purification of correctly oxidized MHC class I heavy-chain molecules under denaturing conditions: a

novel strategy exploiting disulfide assisted protein folding. Protein Sci. 12, 551-559.

Frankild S., de Boer R.J., Lund O., Nielsen M. & Kesmir C. (2008) Amino acid similarity accounts for T cell

cross-reactivity and for "holes" in the T cell repertoire. PLoS. One. 3, e1831.

Gao M., Zhang R., Li M., Li S., Cao Y., Ma B. & Wang J. (2012) An ELISA based on the repeated foot-and-

mouth disease virus 3B epitope peptide can distinguish infected and vaccinated cattle. Appl. Microbiol.

Biotechnol. 93, 1271-1279.

Gay D., Saunders T., Camper S. & Weigert M. (1993) Receptor editing: an approach by autoreactive B

cells to escape tolerance. J. Exp. Med. 177, 999-1008.

Germain R.N. (2012) Maintaining system homeostasis: the third law of Newtonian immunology. Nat.

Immunol. 13, 902-906.

Golde W.T., Pacheco J.M., Duque H., Doel T., Penfold B., Ferman G.S., Gregg D.R. & Rodriguez L.L. (2005)

Vaccination against foot-and-mouth disease virus confers complete clinical protection in 7 days and

partial protection in 4 days: Use in emergency outbreak response. Vaccine 23, 5775-5782.

Goldszmid R.S., Trinchieri G. (2012) The price of immunity. Nat. Immunol. 13, 932-938.

Gorer P.A., Lyman S. & Snell G.D. (1948) Studies on the Genetic and Antigenic Basis of Tumour

Transplantation. Linkage between a Histocompatibility Gene and 'Fused' in Mice. Proc. R. Soc. Lond. B.

135, 499-505.

Grey H.M., Kubo R.T., Colon S.M., Poulik M.D., Cresswell P., Springer T., Turner M. & Strominger J.L.

(1973) The small subunit of HL-A antigens is beta 2-microglobulin. J. Exp. Med. 138, 1608-1612.

66

Grubman M.J., Baxt B. (2004) Foot-and-mouth disease. Clin. Microbiol. Rev. 17, 465-493.

Grubman M.J., Moraes M., Schutta C., Barrera J., Neilan J., Ettyreddy D., Butman B.T., Brough D.E. &

Brake D.A. (2010) Adenovirus serotype 5-vectored foot-and-mouth disease subunit vaccines: the first

decade. Future Virol. 5:51-64.,

Grubman M.J., Zellner M., Bablanian G., Mason P.W. & Piccone M.E. (1995) Identification of the active-

site residues of the 3C proteinase of foot-and-mouth disease virus. Virology 213, 581-589.

Guzman E., Taylor G., Charleston B., Skinner M.A. & Ellis S.A. (2008) An MHC-restricted CD8+ T-cell

response is induced in cattle by foot-and-mouth disease virus (FMDV) infection and also following

vaccination with inactivated FMDV. J. Gen. Virol. 89, 667-675.

Harndahl M., Justesen S., Lamberth K., Roder G., Nielsen M. & Buus S. (2009) Peptide binding to HLA

class I molecules: homogenous, high-throughput screening, and affinity assays. J. Biomol. Screen. 14,

173-180.

Harndahl M., Rasmussen M., Roder G., Dalgaard P., I, Sorensen M., Nielsen M. & Buus S. (2012) Peptide-

MHC class I stability is a better predictor than peptide affinity of CTL immunogenicity. Eur. J. Immunol.

42, 1405-1416.

Hart H.E., Greenwald E.B. (1979a) Scintillation proximity assay (SPA)--a new method of immunoassay.

Direct and inhibition mode detection with human albumin and rabbit antihuman albumin. Mol.

Immunol. 16, 265-267.

Hart H.E., Greenwald E.B. (1979b) Scintillation-proximity assay of antigen-antibody binding kinetics:

concise communication. J. Nucl. Med. 20, 1062-1065.

Hartley S.B., Cooke M.P., Fulcher D.A., Harris A.W., Cory S., Basten A. & Goodnow C.C. (1993) Elimination

of self-reactive B lymphocytes proceeds in two stages: arrested development and cell death. Cell 72,

325-335.

Harty J.T., Tvinnereim A.R. & White D.W. (2000) CD8+ T cell effector mechanisms in resistance to

infection. Annu. Rev. Immunol. 18, 275-308.

He X.S., Rehermann B., Lopez-Labrador F.X., Boisvert J., Cheung R., Mumm J., Wedemeyer H., Berenguer

M., Wright T.L., Davis M.M. & Greenberg H.B. (1999) Quantitative analysis of hepatitis C virus-specific

CD8(+) T cells in peripheral blood and liver using peptide-MHC tetramers. Proc. Natl. Acad. Sci. U. S. A

96, 5692-5697.

Heegaard P.M., Dedieu L., Johnson N., Le Potier M.F., Mockey M., Mutinelli F., Vahlenkamp T., Vascellari

M. & Sorensen N.S. (2011) Adjuvants and delivery systems in veterinary vaccinology: current state and

future developments. Arch. Virol. 156, 183-202.

Ho C.S., Ando A., Essler S.E., Rogel-Gaillard C., Lee J.H., Lunney J.K., Schook L. & Smith D.M. The Swine

Leukocyte Antigen (SLA) Nomenclature System: Current Status after 10 Years of its Establishment.

Conference of the International Society for Animal Genetics. [33rd.]. 20-7-2012.

Ref Type: Conference Proceeding

67

Ho C.S., Franzo-Romain M.H., Lee Y.J., Lee J.H. & Smith D.M. (2009a) Sequence-based characterization of

swine leucocyte antigen alleles in commercially available porcine cell lines. Int. J. Immunogenet. 36, 231-

234.

Ho C.S., Lunney J.K., Ando A., Rogel-Gaillard C., Lee J.H., Schook L.B. & Smith D.M. (2009b)

Nomenclature for factors of the SLA system, update 2008. Tissue Antigens 73, 307-315.

Ho C.S., Lunney J.K., Franzo-Romain M.H., Martens G.W., Lee Y.J., Lee J.H., Wysocki M., Rowland R.R. &

Smith D.M. (2009c) Molecular characterization of swine leucocyte antigen class I genes in outbred pig

populations. Anim Genet. 40, 468-478.

Ho C.S., Lunney J.K., Lee J.H., Franzo-Romain M.H., Martens G.W., Rowland R.R. & Smith D.M. (2010a)

Molecular characterization of swine leucocyte antigen class II genes in outbred pig populations. Anim

Genet. 41, 428-432.

Ho C.S., Martens G.W., Amoss M.S., Jr., Gomez-Raya L., Beattie C.W. & Smith D.M. (2010b) Swine

leukocyte antigen (SLA) diversity in Sinclair and Hanford swine. Dev. Comp Immunol. 34, 250-257.

Ho C.S., Rochelle E.S., Martens G.W., Schook L.B. & Smith D.M. (2006) Characterization of swine

leukocyte antigen polymorphism by sequence-based and PCR-SSP methods in Meishan pigs.

Immunogenetics 58, 873-882.

Hoof I., Peters B., Sidney J., Pedersen L.E., Sette A., Lund O., Buus S. & Nielsen M. (2009) NetMHCpan, a

method for MHC class I binding prediction beyond humans. Immunogenetics 61, 1-13.

Horton R., Wilming L., Rand V., Lovering R.C., Bruford E.A., Khodiyar V.K., Lush M.J., Povey S., Talbot C.C.,

Jr., Wright M.W., Wain H.M., Trowsdale J., Ziegler A. & Beck S. (2004) Gene map of the extended human

MHC. Nat. Rev. Genet. 5, 889-899.

Hunt D.F., Henderson R.A., Shabanowitz J., Sakaguchi K., Michel H., Sevilir N., Cox A.L., Appella E. &

Engelhard V.H. (1992) Characterization of peptides bound to the class I MHC molecule HLA-A2.1 by mass

spectrometry. Science 255, 1261-1263.

Hutchings P.R., Cambridge G., Tite J.P., Meager T. & Cooke A. (1989) The detection and enumeration of

cytokine-secreting cells in mice and man and the clinical application of these assays. J. Immunol.

Methods 120, 1-8.

Janeway C.A., Jr. (1989) Approaching the asymptote? Evolution and revolution in immunology. Cold

Spring Harb. Symp. Quant. Biol. 54 Pt 1, 1-13.

Jegaskanda S., Reece J.C., De R.R., Stambas J., Sullivan L., Brooks A.G., Kent S.J. & Sexton A. (2012)

Comparison of influenza and SIV specific CD8 T cell responses in macaques. PLoS. One. 7, e32431.

Jo S.K., Kim H.S., Cho S.W. & Seo S.H. (2007) Pathogenesis and inflammatory responses of swine H1N2

influenza viruses in pigs. Virus Res. 129, 64-70.

Kelley J., Walter L. & Trowsdale J. (2005) Comparative genomics of major histocompatibility complexes.

Immunogenetics 56, 683-695.

Kida H., Ito T., Yasuda J., Shimizu Y., Itakura C., Shortridge K.F., Kawaoka Y. & Webster R.G. (1994)

Potential for transmission of avian influenza viruses to pigs. J. Gen. Virol. 75 ( Pt 9), 2183-2188.

68

Kiepiela P., Leslie A.J., Honeyborne I., Ramduth D., Thobakgale C., Chetty S., Rathnavalu P., Moore C.,

Pfafferott K.J., Hilton L., Zimbwa P., Moore S., Allen T., Brander C., Addo M.M., Altfeld M., James I.,

Mallal S., Bunce M., Barber L.D., Szinger J., Day C., Klenerman P., Mullins J., Korber B., Coovadia H.M.,

Walker B.D. & Goulder P.J. (2004) Dominant influence of HLA-B in mediating the potential co-evolution

of HIV and HLA. Nature 432, 769-775.

Kisielow P., Bluthmann H., Staerz U.D., Steinmetz M. & von B.H. (1988) Tolerance in T-cell-receptor

transgenic mice involves deletion of nonmature CD4+8+ thymocytes. Nature 333, 742-746.

Klenerman P., Cerundolo V. & Dunbar P.R. (2002) Tracking T cells with tetramers: new tales from new

tools. Nat. Rev. Immunol. 2, 263-272.

Kristensen B. (1988) Current Status of SLA Class I and II Serology. In: Improving Genetic Disease

Resistance in Farm Animals. Kluwer Academic Publishers., Brussels, Belgium, pp 23-32

Kubo R.T., Sette A., Grey H.M., Appella E., Sakaguchi K., Zhu N.Z., Arnott D., Sherman N., Shabanowitz J.,

Michel H. & . (1994) Definition of specific peptide motifs for four major HLA-A alleles. J. Immunol. 152,

3913-3924.

Kundin W.D. (1970) Hong Kong A-2 influenza virus infection among swine during a human epidemic in

Taiwan. Nature 228, 857.

Kusza S., Flori L., Gao Y., Teillaud A., Hu R., Lemonnier G., Bosze Z., Bourneuf E., Vincent-Naulleau S. &

Rogel-Gaillard C. (2011) Transcription specificity of the class Ib genes SLA-6, SLA-7 and SLA-8 of the

swine major histocompatibility complex and comparison with class Ia genes. Anim Genet. 42, 510-520.

Lacey C., Wilkie B.N., Kennedy B.W. & Mallard B.A. (1989) Genetic and other effects on bacterial

phagocytosis and killing by cultured peripheral blood monocytes of SLA-defined miniature pigs. Anim

Genet. 20, 371-381.

Larsen M.V., Lundegaard C., Lamberth K., Buus S., Brunak S., Lund O. & Nielsen M. (2005) An integrative

approach to CTL epitope prediction: a combined algorithm integrating MHC class I binding, TAP

transport efficiency, and proteasomal cleavage predictions. Eur. J. Immunol. 35, 2295-2303.

Leisner C., Loeth N., Lamberth K., Justesen S., Sylvester-Hvid C., Schmidt E.G., Claesson M., Buus S. &

Stryhn A. (2008) One-pot, mix-and-read peptide-MHC tetramers. PLoS. One. 3, e1678.

Lemke C.D., Graham J.B., Lubaroff D.M. & Salem A.K. (2011) Development of an MHC class I L(d)-

restricted PSA peptide-loaded tetramer for detection of PSA-specific CD8+ T cells in the mouse. Prostate

Cancer Prostatic. Dis. 14, 118-121.

Li X., Yang X., Jiang Y. & Liu J. (2005) A novel HBV DNA vaccine based on T cell epitopes and its potential

therapeutic effect in HBV transgenic mice. Int. Immunol. 17, 1293-1302.

Lumsden J.S., Kennedy B.W., Mallard B.A. & Wilkie B.N. (1993) The influence of the swine major

histocompatibility genes on antibody and cell-mediated immune responses to immunization with an

aromatic-dependent mutant of Salmonella typhimurium. Can. J. Vet. Res. 57, 14-18.

Lundegaard C., Lund O. & Nielsen M. (2011) Prediction of epitopes using neural network based methods.

J. Immunol. Methods 374, 26-34.

69

Lunney J.K., Ho C.S., Wysocki M. & Smith D.M. (2009) Molecular genetics of the swine major

histocompatibility complex, the SLA complex. Dev. Comp Immunol. 33, 362-374.

Madden D.R. (1995) The three-dimensional structure of peptide-MHC complexes. Annu. Rev. Immunol.

13, 587-622.

Madden D.R., Gorga J.C., Strominger J.L. & Wiley D.C. (1992) The three-dimensional structure of HLA-

B27 at 2.1 A resolution suggests a general mechanism for tight peptide binding to MHC. Cell 70, 1035-

1048.

Madden K.B., Murrell K.D. & Lunney J.K. (1990) Trichinella spiralis: major histocompatibility complex-

associated elimination of encysted muscle larvae in swine. Exp. Parasitol. 70, 443-451.

Mallard B.A., Wilkie B.N. & Kennedy B.W. (1989a) Influence of major histocompatibility genes on serum

hemolytic complement activity in miniature swine. Am. J. Vet. Res. 50, 359-363.

Mallard B.A., Wilkie B.N. & Kennedy B.W. (1989b) The influence of the swine major histocompatibility

genes (SLA) on variation in serum immunoglobulin (Ig) concentration. Vet. Immunol. Immunopathol. 21,

139-151.

Malyguine A., Strobl S., Zaritskaya L., Baseler M. & Shafer-Weaver K. (2007) New approaches for

monitoring CTL activity in clinical trials. Adv. Exp. Med. Biol. 601, 273-284.

Martens G.W., Lunney J.K., Baker J.E. & Smith D.M. (2003) Rapid assignment of swine leukocyte antigen

haplotypes in pedigreed herds using a polymerase chain reaction-based assay. Immunogenetics 55, 395-

401.

Matsumura M., Fremont D.H., Peterson P.A. & Wilson I.A. (1992) Emerging principles for the recognition

of peptide antigens by MHC class I molecules. Science 257, 927-934.

Mayr G.A., Chinsangaram J. & Grubman M.J. (1999) Development of replication-defective adenovirus

serotype 5 containing the capsid and 3C protease coding regions of foot-and-mouth disease virus as a

vaccine candidate. Virology 263, 496-506.

Mayr G.A., O'Donnell V., Chinsangaram J., Mason P.W. & Grubman M.J. (2001) Immune responses and

protection against foot-and-mouth disease virus (FMDV) challenge in swine vaccinated with adenovirus-

FMDV constructs. Vaccine 19, 2152-2162.

Medzhitov R. (2007) Recognition of microorganisms and activation of the immune response. Nature

449, 819-826.

Mescher M.F., Curtsinger J.M., Agarwal P., Casey K.A., Gerner M., Hammerbeck C.D., Popescu F. & Xiao

Z. (2006) Signals required for programming effector and memory development by CD8+ T cells.

Immunol. Rev. 211, 81-92.

Messaoudi I., Estep R., Robinson B. & Wong S.W. (2011) Nonhuman primate models of human

immunology. Antioxid. Redox. Signal. 14, 261-273.

Moraes M.P., Mayr G.A., Mason P.W. & Grubman M.J. (2002) Early protection against homologous

challenge after a single dose of replication-defective human adenovirus type 5 expressing capsid

proteins of foot-and-mouth disease virus (FMDV) strain A24. Vaccine 20, 1631-1639.

70

Mullbacher A., Lobigs M., Yewdell J.W., Bennink J.R., Tha H.R. & Blanden R.V. (1999) High peptide

affinity for MHC class I does not correlate with immunodominance. Scand. J. Immunol. 50, 420-426.

Muller J.D., Wilkins M., Foord A.J., Dolezal O., Yu M., Heine H.G. & Wang L.F. (2010) Improvement of a

recombinant antibody-based serological assay for foot-and-mouth disease virus. J. Immunol. Methods

352, 81-88.

Murphy K., Travers P. & Walport M. (2008) Innate Immunnity. In: Immuno Biology. Garland Science,

Taylor & Francis Group, New York, pp 39-108

Murray P.J., Smale S.T. (2012) Restraint of inflammatory signaling by interdependent strata of negative

regulatory pathways. Nat. Immunol. 13, 916-924.

Nielsen M., Lundegaard C., Blicher T., Lamberth K., Harndahl M., Justesen S., Roder G., Peters B., Sette

A., Lund O. & Buus S. (2007) NetMHCpan, a method for quantitative predictions of peptide binding to

any HLA-A and -B locus protein of known sequence. PLoS. One. 2, e796.

Nordly P., Rose F., Christensen D., Nielsen H.M., Andersen P., Agger E.M. & Foged C. (2011) Immunity by

formulation design: induction of high CD8+ T-cell responses by poly(I:C) incorporated into the CAF01

adjuvant via a double emulsion method. J. Control Release 150, 307-317.

Norimine J., Han S. & Brown W.C. (2006) Quantitation of Anaplasma marginale major surface protein

(MSP)1a and MSP2 epitope-specific CD4+ T lymphocytes using bovine DRB3*1101 and DRB3*1201

tetramers. Immunogenetics 58, 726-739.

Ogg G.S., Dunbar P.R., Cerundolo V., McMichael A.J., Lemoine N.R. & Savage P. (2000) Sensitization of

tumour cells to lysis by virus-specific CTL using antibody-targeted MHC class I/peptide complexes. Br. J.

Cancer 82, 1058-1062.

Olsen C.W. (2002) The emergence of novel swine influenza viruses in North America. Virus Res. 85, 199-

210.

Olvera-Gomez I., Hamilton S.E., Xiao Z., Guimaraes C.P., Ploegh H.L., Hogquist K.A., Wang L. & Jameson

S.C. (2012) Cholera toxin activates nonconventional adjuvant pathways that induce protective CD8 T-cell

responses after epicutaneous vaccination. Proc. Natl. Acad. Sci. U. S. A 109, 2072-2077.

Orr H.T., Lancet D., Robb R.J., Lopez de Castro J.A. & Strominger J.L. (1979) The heavy chain of human

histocompatibility antigen HLA-B7 contains an immunoglobulin-like region. Nature 282, 266-270.

Ostergaard P.L., Nissen M.H., Hansen N.J., Nielsen L.L., Lauenmoller S.L., Blicher T., Nansen A., Sylvester-

Hvid C., Thromsen A.R. & Buus S. (2001) Efficient assembly of recombinant major histocompatibility

complex class I molecules with preformed disulfide bonds. Eur. J. Immunol. 31, 2986-2996.

Pacheco J.M., Brum M.C., Moraes M.P., Golde W.T. & Grubman M.J. (2005) Rapid protection of cattle

from direct challenge with foot-and-mouth disease virus (FMDV) by a single inoculation with an

adenovirus-vectored FMDV subunit vaccine. Virology 337, 205-209.

Parish I.A., Kaech S.M. (2009) Diversity in CD8(+) T cell differentiation. Curr. Opin. Immunol. 21, 291-297.

71

Patch J.R., Pedersen L.E., Toka F.N., Moraes M., Grubman M.J., Nielsen M., Jungersen G., Buus S. &

Golde W.T. (2011) Induction of foot-and-mouth disease virus-specific cytotoxic T cell killing by

vaccination. Clin. Vaccine Immunol. 18, 280-288.

Pedersen L.E., Harndahl M., Nielsen M., Patch J.R., Jungersen G., Buus S. & Golde W.T. (2012a)

Identification of peptides from foot-and-mouth disease virus structural proteins bound by class I swine

leukocyte antigen (SLA) alleles, SLA-1*0401 and SLA-2*0401. Anim Genet.

Pedersen L.E., Harndahl M., Rasmussen M., Lamberth K., Golde W.T., Lund O., Nielsen M. & Buus S.

(2011) Porcine major histocompatibility complex (MHC) class I molecules and analysis of their peptide-

binding specificities. Immunogenetics 63, 821-834.

Pedersen L.E., Rasmussen M., Harndahl M., Nielsen M., Buus S. & Jungersen G. (2012b) An analysis of

affinity and stability in the identification of peptides bound by Swine Leukocyte Antigens (SLA)

combining matrix- and NetMHCpan based peptide selection. Immunogenetics Manuscript ready for

submisssion.,

Pedersen L.O., Stryhn A., Holter T.L., Etzerodt M., Gerwien J., Nissen M.H., Thogersen H.C. & Buus S.

(1995) The interaction of beta 2-microglobulin (beta 2m) with mouse class I major histocompatibility

antigens and its ability to support peptide binding. A comparison of human and mouse beta 2m. Eur. J.

Immunol. 25, 1609-1616.

Penn D.J., Damjanovich K. & Potts W.K. (2002) MHC heterozygosity confers a selective advantage against

multiple-strain infections. Proc. Natl. Acad. Sci. U. S. A 99, 11260-11264.

Racaniello V.R. (2006) Picornaviridae: the viruses and their replication. In: Fields Virology, 5th edn.

Lippincott Williams & Wilkins, Philidelphia, PA, pp 795-838.

Renard C., Kristensen B., Gautschi C., Hruban V., Fredholm M. & Vaiman M. (1988) Joint report of the

first international comparison test on swine lymphocyte alloantigens (SLA). Anim Genet. 19, 63-72.

Richt J.A., Lager K.M., Janke B.H., Woods R.D., Webster R.G. & Webby R.J. (2003) Pathogenic and

antigenic properties of phylogenetically distinct reassortant H3N2 swine influenza viruses cocirculating

in the United States. J. Clin. Microbiol. 41, 3198-3205.

Robert B., Guillaume P., Luescher I., Romero P. & Mach J.P. (2000) Antibody-conjugated MHC class I

tetramers can target tumor cells for specific lysis by T lymphocytes. Eur. J. Immunol. 30, 3165-3170.

Robins H.S., Campregher P.V., Srivastava S.K., Wacher A., Turtle C.J., Kahsai O., Riddell S.R., Warren E.H.

& Carlson C.S. (2009) Comprehensive assessment of T-cell receptor beta-chain diversity in alphabeta T

cells. Blood 114, 4099-4107.

Rodriguez L.L., Grubman M.J. (2009) Foot and mouth disease virus vaccines. Vaccine 27 Suppl 4, D90-

D94.

Romero P., Cerottini J.C. & Speiser D.E. (2004) Monitoring tumor antigen specific T-cell responses in

cancer patients and phase I clinical trials of peptide-based vaccination. Cancer Immunol. Immunother.

53, 249-255.

Romero P., Dunbar P.R., Valmori D., Pittet M., Ogg G.S., Rimoldi D., Chen J.L., Lienard D., Cerottini J.C. &

Cerundolo V. (1998) Ex vivo staining of metastatic lymph nodes by class I major histocompatibility

72

complex tetramers reveals high numbers of antigen-experienced tumor-specific cytolytic T lymphocytes.

J. Exp. Med. 188, 1641-1650.

Rothschild M.F., Chen H.L., Christian L.L., Lie W.R., Venier L., Cooper M., Briggs C. & Warner C.M. (1984)

Breed and swine lymphocyte antigen haplotype differences in agglutination titers following vaccination

with B. bronchiseptica. J. Anim Sci. 59, 643-649.

Rueckert R.R., Wimmer E. (1984) Systematic nomenclature of picornavirus proteins. J. Virol. 50, 957-959.

Saalmuller A. (2006) New understanding of immunological mechanisms. Vet. Microbiol. 117, 32-38.

Salter R.D., Norment A.M., Chen B.P., Clayberger C., Krensky A.M., Littman D.R. & Parham P. (1989)

Polymorphism in the alpha 3 domain of HLA-A molecules affects binding to CD8. Nature 338, 345-347.

Sanz-Parra A., Sobrino F. & Ley V. (1998) Infection with foot-and-mouth disease virus results in a rapid

reduction of MHC class I surface expression. J. Gen. Virol. 79 ( Pt 3), 433-436.

Saxena R.K., Tripathi P. & Rawat G. (2012) Pandemism of swine flu and its prospective drug therapy. Eur.

J. Clin. Microbiol. Infect. Dis.

Schatz D.G. (2004) V(D)J recombination. Immunol. Rev. 200, 5-11.

Schiott A., Larsson K., Manniche S., Kalliomaki S., Heydenreich A.V., Dalsgaard K. & Kirkby N. (2011)

Posintro-HBsAg, a modified ISCOM including HBsAg, induces strong cellular and humoral responses. Int.

J. Pharm. 414, 312-320.

Schmidt E.G., Buus S., Thorn M., Stryhn A., Leisner C. & Claesson M.H. (2009) Peptide specific expansion

of CD8(+) T cells by recombinant plate bound MHC/peptide complexes. J. Immunol. Methods 340, 25-32.

Scudamore J.M., Harris D.M. (2002) Control of foot and mouth disease: lessons from the experience of

the outbreak in Great Britain in 2001. Rev. Sci. Tech. 21, 699-710.

Seder R.A., Darrah P.A. & Roederer M. (2008) T-cell quality in memory and protection: implications for

vaccine design. Nat. Rev. Immunol. 8, 247-258.

Shapiro-Shelef M., Calame K. (2005) Regulation of plasma-cell development. Nat. Rev. Immunol. 5, 230-

242.

Singer D.S., Mozes E., Kirshner S. & Kohn L.D. (1997) Role of MHC class I molecules in autoimmune

disease. Crit Rev. Immunol. 17, 463-468.

Smith D.M., Lunney J.K., Martens G.W., Ando A., Lee J.H., Ho C.S., Schook L., Renard C. & Chardon P.

(2005) Nomenclature for factors of the SLA class-I system, 2004. Tissue Antigens 65, 136-149.

Sorensen K.J., de S.K., Dyrting K.C., Grazioli S. & Haas B. (2005) Differentiation of foot-and-mouth

disease virus infected animals from vaccinated animals using a blocking ELISA based on baculovirus

expressed FMDV 3ABC antigen and a 3ABC monoclonal antibody. Arch. Virol. 150, 805-814.

Stanway G., Brown F., Christian P., Hovi T., Hyypiä T., King A.M.Q., Knowles N.J., Lemon S.M., Minor P.D.,

Pallansch M.A., Palmenberg A.C. & Skern T. (2005) Virus taxonomy: classification and nomenclature of

73

viruses: eighth report of the International Committee on Taxonomy of Viruses. Elsevier Academic Press,

San Diego, CA, pp 757-778.

Stryhn A. Yellow fever peptides identified to be epitopes in some donors but not others, despite sharing

a certain HLA allele susceptible for peptide binding. Only very few epitopes are dominant in all donors.

27-11-2012.

Ref Type: Personal Communication

Stryhn A., Pedersen L.O., Romme T., Holm C.B., Holm A. & Buus S. (1996) Peptide binding specificity of

major histocompatibility complex class I resolved into an array of apparently independent

subspecificities: quantitation by peptide libraries and improved prediction of binding. Eur. J. Immunol.

26, 1911-1918.

Suhrbier A. (1997) Multi-epitope DNA vaccines. Immunol. Cell Biol. 75, 402-408.

Sylvester-Hvid C., Kristensen N., Blicher T., Ferre H., Lauemoller S.L., Wolf X.A., Lamberth K., Nissen

M.H., Pedersen L.O. & Buus S. (2002) Establishment of a quantitative ELISA capable of determining

peptide - MHC class I interaction. Tissue Antigens 59, 251-258.

Tanaka-Matsuda M., Ando A., Rogel-Gaillard C., Chardon P. & Uenishi H. (2009) Difference in number of

loci of swine leukocyte antigen classical class I genes among haplotypes. Genomics 93, 261-273.

Thakur A., Pedersen L.E. & Jungersen G. (2012) Immune markers and correlates of protection for vaccine

induced immune responses. Vaccine 30, 4907-4920.

Thompson D., Muriel P., Russell D., Osborne P., Bromley A., Rowland M., Creigh-Tyte S. & Brown C.

(2002) Economic costs of the foot and mouth disease outbreak in the United Kingdom in 2001. Rev. Sci.

Tech. 21, 675-687.

Thomson S.A., Sherritt M.A., Medveczky J., Elliott S.L., Moss D.J., Fernando G.J., Brown L.E. & Suhrbier A.

(1998) Delivery of multiple CD8 cytotoxic T cell epitopes by DNA vaccination. J. Immunol. 160, 1717-

1723.

thor Straten P., Kirkin A.F., Siim E., Dahlstrom K., Drzewiecki K.T., Seremet T., Zeuthen J., Becker J.C. &

Guldberg P. (2000) Tumor infiltrating lymphocytes in melanoma comprise high numbers of T-cell

clonotypes that are lost during in vitro culture. Clin. Immunol. 96, 94-99.

Tonegawa S. (1983) Somatic generation of antibody diversity. Nature 302, 575-581.

Torremorell M., Allerson M., Corzo C., Diaz A. & Gramer M. (2012) Transmission of Influenza A Virus in

Pigs. Transbound. Emerg. Dis.

Trebbien R., Larsen L.E. & Viuff B.M. (2011) Distribution of sialic acid receptors and influenza A virus of

avian and swine origin in experimentally infected pigs. Virol. J. 8, 434.

Ujvari B., Belov K. (2011) Major histocompatibility complex (MHC) markers in conservation biology. Int.

J. Mol. Sci. 12, 5168-5186.

Ullman E.F., Kirakossian H., Singh S., Wu Z.P., Irvin B.R., Pease J.S., Switchenko A.C., Irvine J.D., Dafforn

A., Skold C.N. & . (1994) Luminescent oxygen channeling immunoassay: measurement of particle binding

kinetics by chemiluminescence. Proc. Natl. Acad. Sci. U. S. A 91, 5426-5430.

74

Ullman E.F., Kirakossian H., Switchenko A.C., Ishkanian J., Ericson M., Wartchow C.A., Pirio M., Pease J.,

Irvin B.R., Singh S., Singh R., Patel R., Dafforn A., Davalian D., Skold C., Kurn N. & Wagner D.B. (1996)

Luminescent oxygen channeling assay (LOCI): sensitive, broadly applicable homogeneous immunoassay

method. Clin. Chem. 42, 1518-1526.

Vaiman M., Renard C., LaFage P., Ameteau J. & Nizza P. (1970) Evidence for a histocompatibility system

in swine (SL-A). Transplantation 10, 155-164.

van der Burg S.H., Visseren M.J., Brandt R.M., Kast W.M. & Melief C.J. (1996) Immunogenicity of

peptides bound to MHC class I molecules depends on the MHC-peptide complex stability. J. Immunol.

156, 3308-3314.

Van Weemen B.K., Schuurs A.H. (1971) Immunoassay using antigen-enzyme conjugates. FEBS Lett. 15,

232-236.

Viza D., Sugar J.R. & Binns R.M. (1970) Lymphocyte stimulation in pigs: evidence for the existence of a

single major histocompatibility locus, PL-A. Nature 227, 949-950.

Webster R.G., Bean W.J., Gorman O.T., Chambers T.M. & Kawaoka Y. (1992) Evolution and ecology of

influenza A viruses. Microbiol. Rev. 56, 152-179.

Wei H., Wang R., Yuan Z., Chen C.Y., Huang D., Halliday L., Zhong W., Zeng G., Shen Y., Shen L., Wang Y.

& Chen Z.W. (2009) DR*W201/P65 tetramer visualization of epitope-specific CD4 T-cell during M.

tuberculosis infection and its resting memory pool after BCG vaccination. PLoS. One. 4, e6905.

Wijdeven R.H., Bakker J.M., Paul P. & Neefjes J. (2012) Exploring genome-wide datasets of MHC class II

antigen presentation. Mol. Immunol.

Wooldridge L., Lissina A., Cole D.K., van den Berg H.A., Price D.A. & Sewell A.K. (2009) Tricks with

tetramers: how to get the most from multimeric peptide-MHC. Immunology 126, 147-164.

Xu X.N., Screaton G.R. (2002) MHC/peptide tetramer-based studies of T cell function. J. Immunol.

Methods 268, 21-28.

Zhang N., Qi J., Feng S., Gao F., Liu J., Pan X., Chen R., Li Q., Chen Z., Li X., Xia C. & Gao G.F. (2011) Crystal

structure of swine major histocompatibility complex class I SLA-1 0401 and identification of 2009

pandemic swine-origin influenza A H1N1 virus cytotoxic T lymphocyte epitope peptides. J. Virol. 85,

11709-11724.

Zuber B., Levitsky V., Jonsson G., Paulie S., Samarina A., Grundstrom S., Metkar S., Norell H., Callender

G.G., Froelich C. & Ahlborg N. (2005) Detection of human perforin by ELISpot and ELISA: ex vivo

identification of virus-specific cells. J. Immunol. Methods 302, 13-25.

PhD thesis Lasse Eggers Pedersen

DTU Veterinærinstituttet

Danmarks Tekniske Universitet Bülowsvej 27

1870 Frederiksberg C.

Tlf. 35 88 60 00

www.vet.dtu.dk

Analysis of Swine Leukocyte Antigen Peptide …...Analysis of Swine Leukocyte Antigen Peptide...

Documents

Transcript of Analysis of Swine Leukocyte Antigen Peptide …...Analysis of Swine Leukocyte Antigen Peptide...